Wind turbine variable pitch system backup power charging and discharging monitoring method and system

Abstract

The present disclosure provides a charge-discharge monitoring method and system for a backup power supply of a wind turbine variable pitch system, which actively injects a short and transient micro-pulse current into a supercapacitor when the wind turbine is in a standby or other preset safe state, thereby creating dynamic data for analysis without affecting the backup capability of the system. By high-frequency acquisition and analysis of the instantaneous voltage and current response caused by the pulse, the original equivalent series resistance and original capacitance reflecting the health status of the supercapacitor can be identified. In order to eliminate the interference of the environmental temperature on the measurement results, firstly, the identified original parameters are normalized by using a temperature compensation model; then, the normalized parameter sequence is optimized by using a Kalman filter to obtain reliable aging state estimation values. Finally, based on the optimized resistance and capacitance states, the most real health status is comprehensively evaluated and a warning is issued, realizing high-precision predictive diagnosis of early faults of the backup power supply.

Description

Technical Field

[0001] This invention relates to the field of new energy wind power generation, and in particular to a method and system for monitoring the charging and discharging of backup power supply for wind turbine pitch control systems. Background Technology

[0002] Wind turbine generators are core equipment in the renewable energy sector. Their pitch control system adjusts the blade pitch angle to control the windward area of ​​the rotor, thereby achieving precise control over the generator's output power and speed. In emergencies, feathering operation ensures the generator's safety. To ensure the pitch control system can still perform critical safety feathering actions under abnormal conditions such as grid outages or main power failures, an independent backup power supply is typically provided. With technological advancements, supercapacitors, due to their high power density, extremely wide operating temperature range, and ultra-long cycle life, are increasingly becoming the mainstream choice for backup power supplies in wind turbine pitch control systems. Their operational reliability directly impacts the asset and operational safety of the entire wind turbine generator set.

[0003] Accurate online SOH assessment of supercapacitor backup power supplies operating in a long-term float-charge standby state remains a significant technical challenge. The core technical issue lies in the fact that backup power supplies spend most of their lifespan in a fully charged or near-fully charged quiescent float-charge state, rarely experiencing deep discharge processes that reflect their internal state. This results in extremely scarce effective dynamic data for analyzing their health status. Existing monitoring methods have limitations: offline capacity testing or internal resistance detection, while accurate, requires shutdown and is invasive, potentially impacting power supply lifespan and leaving the system without backup power during testing; while traditional online voltage and current monitoring methods struggle to accurately assess the SOH of supercapacitors, which is jointly determined by the increase in equivalent series resistance and the decrease in capacitance. Furthermore, backup power supplies installed in wheel hubs operate in harsh environments, and severe temperature fluctuations significantly interfere with ESR and capacitance measurements. Existing online monitoring technologies often fail to effectively decouple temperature effects from actual aging degradation, leading to inaccurate SOH assessments and hindering reliable data for operational and maintenance decisions.

[0004] Therefore, an optimized scheme for monitoring the charging and discharging of the backup power supply for wind turbine pitch systems is desired. Summary of the Invention

[0005] The present invention aims to at least solve one of the technical problems existing in the prior art, and provides a method and system for monitoring the charging and discharging of backup power supply for wind turbine pitch systems.

[0006] In a first aspect, embodiments of the present invention provide a method for monitoring the charging and discharging of a backup power supply for a wind turbine pitch system, comprising: Test triggering and condition checking are performed based on the wind turbine status and charge status to obtain pulse injection commands; In response to the pulse injection command being true, the charging module performs micro-pulse injection and high-frequency data acquisition on the supercapacitor to obtain a raw dataset. Each data point in the raw dataset includes the voltage across the supercapacitor, the current flowing through it, and the surface temperature. Instantaneous identification of key electrical parameters is performed based on the original dataset to obtain the original equivalent series resistance, original capacitance, and average temperature. The original equivalent series resistance, original capacitance, average temperature and the state estimate of the previous time step are normalized and optimized by Kalman filtering to obtain the state estimate of the current time step. Health status assessment and early warning are performed based on the current state estimate to obtain the final health status and health alarm signal.

[0007] Secondly, embodiments of the present invention provide a charging and discharging monitoring system for a backup power supply of a wind turbine pitch system, comprising: The test triggering and condition checking module is used to perform test triggering and condition checking based on the wind turbine status and charge status to obtain pulse injection commands; The micro-pulse injection and high-frequency data acquisition module is used to respond to the pulse injection command being true, and the charging module performs micro-pulse injection and high-frequency data acquisition on the supercapacitor to obtain the original dataset. Each data point in the original dataset includes the voltage across the supercapacitor, the current flowing through it, and the surface temperature. The raw data acquisition module is used to perform instantaneous identification of key electrical parameters based on the raw dataset to obtain the raw equivalent series resistance, raw capacitance, and average temperature. The parameter normalization and Kalman filter optimization module is used to perform parameter normalization and Kalman filter optimization on the original equivalent series resistance, original capacitance, average temperature and the state estimate of the previous time step to obtain the state estimate of the current time step. The health status assessment and early warning module is used to assess and warn of health status based on the current status estimate to obtain the final health status and health alarm signal.

[0008] Compared with existing technologies, this invention proposes a charging and discharging monitoring method for the backup power supply of a wind turbine pitch system. This method actively injects a brief micro-pulse current into the supercapacitor when the wind turbine is in a preset safe state such as standby, thereby creating analyzable dynamic data without affecting the system's backup capability. By acquiring and analyzing the instantaneous voltage and current responses caused by this pulse at high frequency, two core indicators reflecting the health status of the supercapacitor can be immediately identified: the original equivalent series resistance and the original capacitance. To eliminate the interference of ambient temperature on the measurement results, this method first normalizes the identified original parameters using a temperature compensation model to remove the temperature effect; then, a Kalman filter is used to optimize the normalized parameter sequence, filtering out random measurement noise and obtaining a smooth and reliable aging state estimate. Finally, based on the optimized resistance and capacitance states, the most accurate health status is comprehensively evaluated and an early warning is issued, achieving high-precision predictive diagnosis of early backup power supply failures. Attached Figure Description

[0009] The above and other objects, features, and advantages of the present invention will become more apparent from the more detailed description of the embodiments of the invention in conjunction with the accompanying drawings. The drawings are provided to further illustrate the embodiments of the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings, the same reference numerals generally represent the same parts or steps.

[0010] Figure 1 A flowchart illustrating a method for monitoring the charging and discharging of a backup power supply for a wind turbine pitch system according to an embodiment of the present invention; Figure 2 A schematic diagram of the data flow of a method for monitoring the charging and discharging of a backup power supply for a wind turbine pitch system according to an embodiment of the present invention; Figure 3 This is a flowchart illustrating the instantaneous identification of key electrical parameters based on the original dataset to obtain the original equivalent series resistance, original capacitance, and average temperature in the charging and discharging monitoring method of the backup power supply of the wind turbine pitch system according to an embodiment of the present invention. Figure 4 This is a flowchart illustrating the charging and discharging monitoring method for the backup power supply of a wind turbine pitch system according to an embodiment of the present invention, which performs parameter normalization and Kalman filtering optimization on the original equivalent series resistance, original capacitance, average temperature, and the state estimate of the previous moment to obtain the state estimate of the current moment. Figure 5 This is a flowchart illustrating the charging and discharging monitoring method for the backup power supply of a wind turbine pitch system according to an embodiment of the present invention, which performs health status assessment and early warning based on the current state estimate to obtain the final health status and health alarm signal. Figure 6This is a block diagram of a charging and discharging monitoring system for a backup power supply of a wind turbine pitch system according to an embodiment of the present invention. Detailed Implementation

[0011] To enable those skilled in the art to better understand the technical solutions of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0012] Unless otherwise specifically stated, the technical or scientific terms used in the embodiments of this invention should be understood in their ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains. The terms "comprising" or "including," as used in the embodiments of this invention, do not limit the shapes, numbers, steps, actions, operations, components, elements, and / or groups thereof mentioned, nor do they exclude the appearance or addition of one or more other different shapes, numbers, steps, actions, operations, components, elements, and / or groups thereof, or the inclusion of these.

[0013] Unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of the invention. It should also be understood that, for ease of description, the dimensions of the various parts shown in the drawings are not drawn to actual scale, and techniques, methods, and apparatus known to those skilled in the art may not be discussed in detail; however, where appropriate, the illustrated techniques, methods, and apparatus should be considered part of the specification. In all the examples shown and discussed herein, any other specific example may have different values. It should be noted that similar symbols and letters in the following figures denote similar items; therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0014] In the description of the embodiments of the present invention, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In the embodiments of the present invention, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in a suitable manner in any one or more embodiments or examples. Furthermore, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in the embodiments of the present invention, as well as the features of different embodiments or examples.

[0015] Hereinafter, exemplary embodiments according to the present invention will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of the present invention, and not all embodiments of the present invention. It should be understood that the present invention is not limited to the exemplary embodiments described herein.

[0016] The backup power supply of a wind turbine pitch system suffers from sparse effective dynamic data for health assessment due to its prolonged float charging state. Furthermore, drastic temperature changes within the hub severely interfere with the accuracy of monitoring parameters, making it difficult for existing technologies to accurately assess its health status online. Therefore, this application proposes a charging and discharging monitoring method for the backup power supply of a wind turbine pitch system. Specifically, this method first determines the current operating state of the wind turbine and the state of charge of the backup power supply to ensure a safe testing window is found without affecting system safety and backup capacity, and then generates a pulse injection command. In response to this command, the system actively injects a small current pulse into the supercapacitor serving as the backup power supply and simultaneously performs high-frequency data acquisition, thereby actively acquiring a set of raw voltage, current, and temperature datasets rich in its internal state information. Next, based on this raw dataset, the system performs instantaneous identification, accurately calculating two core indicators reflecting the health status of the supercapacitor—the original equivalent series resistance and the original capacitance—by analyzing the voltage step before and after the pulse and the slope of the voltage change during the pulse. To overcome environmental interference, this method further normalizes and optimizes the identified raw parameters: First, it uses a preset temperature compensation model to normalize the parameters measured at the current temperature to their values ​​at the standard temperature, thus eliminating the influence of temperature fluctuations. Then, it uses a Kalman filter to fuse the normalized measurement value with the optimal estimate from the previous moment, effectively filtering out random noise from a single measurement and obtaining a smoother and more reliable state estimate for the current moment. Finally, based on the optimized equivalent series resistance and capacitance estimates, the final health status is calculated and compared with a preset threshold to generate a corresponding health alarm signal, thereby achieving high-precision online monitoring and predictive maintenance of the backup power supply's health status.

[0017] Figure 1 This is a flowchart of a method for monitoring the charging and discharging of a backup power supply for a wind turbine pitch system according to an embodiment of the present invention. Figure 2 This is a data flow diagram illustrating the charging and discharging monitoring method for the backup power supply of a wind turbine pitch system according to an embodiment of the present invention. Figure 1 and Figure 2As shown, the charging and discharging monitoring method and system for the backup power supply of a wind turbine pitch system according to an embodiment of the present invention includes the following steps: S100, performing test triggering and condition checking based on the wind turbine state and state of charge to obtain a pulse injection command; S200, in response to the pulse injection command being true, the charging module performs micro-pulse injection and high-frequency data acquisition on the supercapacitor to obtain an original dataset, wherein each data point in the original dataset includes the voltage across the supercapacitor, the current flowing through it, and the surface temperature; S300, performing instantaneous identification of key electrical parameters based on the original dataset to obtain the original equivalent series resistance, the original capacitance, and the average temperature; S400, performing parameter normalization and Kalman filter optimization on the original equivalent series resistance, the original capacitance, the average temperature, and the state estimate of the previous moment to obtain the state estimate of the current moment; S500, performing health status assessment and early warning based on the state estimate of the current moment to obtain the final health status and health alarm signal.

[0018] Specifically, in step S100, test triggering and condition checks are performed based on the wind turbine status and state of charge to obtain the pulse injection command. It should be understood that since the backup power supply of the wind turbine pitch system is a critical component ensuring the safety of the unit, its health status monitoring must be carried out without affecting the normal power generation of the wind turbine, without interfering with grid stability, and without impairing the emergency response capabilities of the backup power supply itself. Therefore, in the technical solution of this application, test triggering and condition checks are performed based on the wind turbine status and state of charge to lock in a safe and effective execution window for subsequent micro-pulse injection proactive diagnosis. This ensures that the implementation of monitoring activities does not pose any risk to the safe operation of the wind turbine unit and guarantees the rigor and periodicity of the test logic.

[0019] More specifically, in a specific example of this application, test triggering and condition checking based on turbine status and state of charge to obtain a pulse injection command includes: determining whether the turbine status belongs to the set of turbine statuses allowed for testing to obtain a first determination result; determining whether the state of charge exceeds a minimum state of charge threshold to obtain a second determination result; generating a system ready flag in response to both the first and second determination results being true; calculating and determining the compliance of the test cycle interval between the current timestamp and the last test timestamp to obtain an interval ready flag; and generating the pulse injection command based on the interval ready flag and the current timestamp. The set of turbine statuses allowed for testing includes standby, idle, and feathering shutdown.

[0020] In other words, more specifically, the process of executing test triggering and condition checks to obtain the pulse injection command unfolds as follows: First, the real-time operating status provided by the wind turbine main control system is acquired and compared with a preset set of wind turbine states that are allowed to be tested. This set includes the wind turbine's static or safe shutdown states when it is not generating power, such as standby, idle, and feathering shutdown. A first judgment result is obtained after the comparison. At the same time, the state of charge data of the backup power supply itself is acquired and it is determined whether it exceeds a preset minimum state of charge threshold, such as 95%, to ensure that the backup power supply still has sufficient emergency energy storage after performing a micro-discharge test, thus obtaining a second judgment result. When both the first and second judgment results are true, a system ready flag is generated, indicating that the wind turbine state and the power supply state at this moment meet the basic prerequisites for testing. Based on this, the current timestamp and the timestamp of the last test completion are read, and the time difference between the two is calculated to determine whether the difference meets the preset test cycle interval requirements, thus obtaining an interval ready flag. Finally, a pulse injection command is generated based on the final state marked as ready at this interval. This command will directly determine whether to start the subsequent micro-pulse injection and data acquisition process.

[0021] Specifically, in step S200, in response to the pulse injection command being true, the charging module performs micro-pulse injection and high-frequency data acquisition on the supercapacitor to obtain a raw dataset. Each data point in the raw dataset includes the voltage across the supercapacitor, the current flowing through it, and its surface temperature. It should be understood that since the supercapacitor backup power supply is in a static floating charge state most of the time, it lacks dynamic electrical signals that can reflect its internal health status, making it impossible to directly assess its performance degradation. Therefore, in the technical solution of this application, in response to the pulse injection command being true, the charging module further performs micro-pulse injection and high-frequency data acquisition on the supercapacitor to actively and non-destructively excite the transient response of the supercapacitor. This allows for the acquisition of a raw, high-resolution dataset containing information on its equivalent series resistance and capacitance, providing the necessary data foundation for subsequent accurate parameter identification.

[0022] More specifically, in a concrete example of this application, the micro-pulse injection and high-frequency data acquisition process is triggered upon receiving a pulse injection command with a true value. The pitch system controller sends a command containing preset pulse parameters to the backup power supply's charging module via the controller area network bus. Based on this command, the internal power switching devices are precisely controlled to apply a brief constant-current discharge pulse to the supercapacitor. Simultaneously, a high-speed data acquisition unit is activated, sampling the voltage across the supercapacitor, the current flowing through it, and the readings of the temperature sensor attached to its surface at a sampling frequency of 1 kHz. The entire data acquisition window covers the period before the pulse begins, during the pulse duration, and a recovery period after the pulse ends, to fully record the step, sag, and relaxation of the voltage under pulse excitation. After acquisition, all timestamped voltage, current, and temperature sampling points are integrated into a structured raw dataset and transmitted to the next step for processing.

[0023] Specifically, in step S300, key electrical parameters are instantaneously identified based on the original dataset to obtain the original equivalent series resistance, original capacitance, and average temperature. It should be understood that since the original dataset obtained after micropulse injection is a series of discrete voltage, current, and temperature time-series data points, it cannot directly quantify the health state of the supercapacitor, which is ultimately determined by its two key physical parameters: equivalent series resistance and capacitance. Therefore, in the technical solution of this application, key electrical parameters are further instantaneously identified based on the original dataset to obtain the original equivalent series resistance, original capacitance, and average temperature, thereby accurately mapping and converting the acquired transient electrical signal response into parameter values ​​that can characterize its intrinsic physical properties. In this way, a low-dimensional, quantified, but real-time snapshot reflecting the current true performance state of the supercapacitor at a specific temperature can be extracted from high-dimensional time-series data, providing a direct and explicit input for subsequent elimination of temperature effects and optimal state estimation.

[0024] Figure 3 This is a flowchart illustrating the instantaneous identification of key electrical parameters based on the original dataset to obtain the original equivalent series resistance, original capacitance, and average temperature in a method for monitoring the charging and discharging of a backup power supply for a wind turbine pitch system according to an embodiment of the present invention. Figure 3As shown, step S300 includes: S310, extracting pulse event features from the original dataset to obtain the pulse start timestamp, pulse end timestamp, pulse current amplitude, instantaneous voltage before the pulse, instantaneous voltage at the beginning of the pulse, instantaneous voltage at the end of the pulse, and temperature sequence during the pulse; S320, calculating the original equivalent series resistance based on the pulse current amplitude, instantaneous voltage before the pulse, and instantaneous voltage at the beginning of the pulse; S330, calculating the original capacitance based on the pulse start timestamp, pulse end timestamp, pulse current amplitude, instantaneous voltage at the beginning of the pulse, and instantaneous voltage at the end of the pulse; S340, calculating the average temperature based on the temperature sequence during the pulse.

[0025] In step S310, pulse event features are extracted from the original dataset to obtain the pulse start timestamp, pulse end timestamp, pulse current amplitude, instantaneous voltage before the pulse, instantaneous voltage at the beginning of the pulse, instantaneous voltage at the end of the pulse, and temperature sequence during the pulse. It should be understood that since the original dataset is a continuous data stream existing in high-frequency time-series form, it contains all information before and after the pulse excitation, while the physical formulas used to calculate the equivalent series resistance and capacitance only require a few key feature values ​​at the instant the pulse event occurs. Therefore, in the technical solution of this application, pulse event features are further extracted from the original dataset to accurately locate and separate the discrete feature points and feature values ​​that define the pulse event from the continuous data stream. This transforms the original, unstructured time-series data into a set of structured key parameters with clear physical meaning, providing direct and effective input for subsequent quantitative calculations of resistance and capacitance.

[0026] More specifically, in a concrete example of this application, the pulse event feature extraction process begins with analyzing the current time series. An algorithm scans the current series, identifying the sampling point where the absolute current value first exceeds a preset threshold, and recording the timestamp of this point as the pulse start timestamp. Next, the scan continues from this start timestamp, identifying the sampling point where the absolute current value first falls below the threshold, and recording this as the pulse end timestamp. After determining the pulse start and end times, all current sampling values ​​within this time interval are averaged to obtain a stable pulse current amplitude. Subsequently, using the determined pulse start and end timestamps, indexing is performed on the voltage time series: the voltage value before the pulse start timestamp is obtained as the instantaneous voltage before the pulse, the voltage value after the pulse start timestamp is obtained as the instantaneous voltage at the beginning of the pulse, and the voltage value corresponding to the pulse end timestamp is obtained as the instantaneous voltage at the end of the pulse. Finally, all temperature readings from the pulse start to the pulse end time series are extracted from the temperature time series to form the pulse period temperature series.

[0027] In step S320, the original equivalent series resistance is calculated based on the pulse current amplitude, the instantaneous voltage before the pulse, and the instantaneous voltage at the beginning of the pulse. This is expressed by the following formula:

[0028]

[0029] in, The instantaneous voltage before the pulse. The voltage at the initial instant of the pulse. This represents the amplitude of the pulse current.

[0030] It should be understood that the instantaneous voltage jump across the supercapacitor at the moment the current pulse is applied is a direct electrical manifestation of its internal equivalent series resistance, a physical characteristic distinct from the subsequent slow voltage drop caused by capacitor discharge. Therefore, in this application's technical solution, the original equivalent series resistance is calculated based on the pulse current amplitude, the instantaneous voltage before the pulse, and the instantaneous voltage at the beginning of the pulse. This allows for a quantitative solution to this purely resistive voltage drop using Ohm's law. The calculation first subtracts the instantaneous voltage at the beginning of the pulse from the instantaneous voltage before the pulse to obtain the voltage drop value entirely caused by the equivalent series resistance. Then, this voltage drop value is divided by the absolute amplitude of the pulse current to calculate the original equivalent series resistance. This allows for the precise separation and quantification of a key intrinsic health parameter of the tested object from complex transient response data, providing the first core indicator for subsequent health status assessment.

[0031] In step S330, the original capacitance is calculated based on the pulse start timestamp, pulse end timestamp, pulse current amplitude, pulse initial instantaneous voltage, and pulse final instantaneous voltage. This is expressed by the following formula:

[0032]

[0033] in, and These are the pulse end timestamp and the pulse start timestamp, respectively. The amplitude of the pulse current. The voltage at the initial instant of the pulse. This is the instantaneous voltage at the end of the pulse.

[0034] It is understandable that, after eliminating the instantaneous voltage drop caused by the equivalent series resistance, the slow decrease in supercapacitor voltage during the pulse duration is a direct manifestation of its pure capacitive characteristics as an energy storage element, and the rate of change of this behavior is directly related to the capacitance. Therefore, in the technical solution of this application, the original capacitance is further calculated based on the pulse start timestamp, pulse end timestamp, pulse current amplitude, pulse initial instantaneous voltage, and pulse final instantaneous voltage to quantitatively solve this process according to the basic charge-voltage relationship of the capacitor. The calculation first multiplies the absolute amplitude of the pulse current by the pulse duration, i.e., the difference between the pulse end and start timestamps, to obtain the total charge transferred during the pulse; then, the voltage change caused entirely by capacitor discharge is obtained by subtracting the pulse final instantaneous voltage from the pulse initial instantaneous voltage; finally, the calculated total charge is divided by this voltage change to obtain the value of the original capacitance. In this way, another key intrinsic health parameter of the tested object, namely energy storage capacity, can be accurately quantified from the transient response data, thus providing a second core indicator for subsequent comprehensive health status assessment.

[0035] In step S340, the average temperature is calculated based on the temperature sequence during the pulse. It should be understood that since the equivalent series resistance and capacitance of a supercapacitor are both highly temperature-sensitive functions, the parameter values ​​measured at different ambient temperatures will differ. This difference can mask the true performance degradation caused by aging. Therefore, in the technical solution of this application, the average temperature is further calculated based on the temperature sequence during the pulse to obtain a single, stable temperature reference value that can represent the thermal conditions throughout the entire micro-pulse test. This provides an accurate temperature input for the subsequent parameter normalization step, thereby ensuring that temperature-induced parameter fluctuations can be effectively distinguished from actual aging degradation.

[0036] More specifically, in one particular example of this application, the calculation of the average temperature is performed based on the temperature sequence extracted during the pulse in the previous step. This calculation process first sums all discrete temperature samples in the temperature sequence. Next, the total number of sampling points in the temperature sequence is counted. Finally, the calculated sum of temperatures is divided by the total number of sampling points to obtain the arithmetic mean temperature during the micropulse test. This calculated single average temperature value will be used as the input parameter for the subsequent temperature compensation model.

[0037] Specifically, in step S400, the original equivalent series resistance, original capacitance, average temperature, and the previous state estimate are normalized and optimized using Kalman filtering to obtain the current state estimate. It should be understood that the original equivalent series resistance and original capacitance values ​​identified in the previous step include fluctuations caused by temperature changes and may also contain random noise from a single measurement. Directly using these instantaneous values ​​cannot accurately track the slow and monotonous long-term aging trend of the supercapacitor. Therefore, in the technical solution of this application, the original equivalent series resistance, original capacitance, average temperature, and the previous state estimate are further normalized and optimized using Kalman filtering to obtain the current state estimate. This systematically removes the interference of ambient temperature and integrates the current measurement value with the historical best estimate. In this way, a smoother and more reliable current state estimate can be generated that eliminates the influence of temperature and suppresses random errors. This estimate can more realistically reflect the irreversible performance degradation of the supercapacitor due to aging.

[0038] like Figure 4 As shown, step S400 includes: S410, performing parameter normalization on the original equivalent series resistance, original capacitance, and average temperature based on a temperature compensation model to obtain a normalized measurement vector; S420, performing state-optimal estimation based on a Kalman filter on the normalized measurement vector, the state estimate of the previous time step, and the covariance matrix of the previous time step to obtain the state estimate of the current time step, wherein the state estimate of the current time step includes the optimized estimate of the equivalent series resistance and the optimized estimate of the capacitance at the current time step.

[0039] In step S410, the original equivalent series resistance, original capacitance, and average temperature are normalized using a temperature compensation model to obtain a normalized measurement vector. It should be understood that the original equivalent series resistance and original capacitance values ​​identified under different ambient temperatures will fluctuate significantly due to temperature effects. This fluctuation can mask the true performance degradation caused by equipment aging, making measurement results under different test conditions incomparable. Therefore, in the technical solution of this application, the original equivalent series resistance, original capacitance, and average temperature are further normalized using a temperature compensation model to obtain a normalized measurement vector. This allows the parameter values ​​measured at the current average temperature to be corrected and converted to equivalent values ​​at a standard reference temperature using a pre-calibrated temperature compensation model. This generates a normalized measurement vector that eliminates the interference of temperature variables. The parameters in this vector truly reflect the performance changes of the supercapacitor due to aging, providing benchmark-compliant, highly reliable input data for subsequent time-series filtering optimization and trend analysis.

[0040] More specifically, in a specific example of this application, parameter normalization of the original equivalent series resistance, original capacitance, and average temperature based on a temperature compensation model to obtain a normalized measurement vector includes: inputting the average temperature into the temperature compensation model function of the supercapacitor to obtain a resistance compensation factor and a capacitance compensation factor; and dividing the original equivalent series resistance and original capacitance by the resistance compensation factor and the capacitance compensation factor, respectively, to obtain the normalized measurement vector.

[0041] Accordingly, the average temperature is input into the temperature compensation model function of the supercapacitor to obtain the resistance compensation factor and the capacitance compensation factor. It should be understood that since the temperature compensation model aims to describe the deterministic relationship between the equivalent series resistance and capacitance and temperature, to use this model for correction, specific correction coefficients must first be calculated based on the current actual temperature conditions. Therefore, in the technical solution of this application, the average temperature is further input into the temperature compensation model function of the supercapacitor to obtain the resistance compensation factor and the capacitance compensation factor, thereby solving for the precise proportional relationship between the actual measured value and the reference value at the standard temperature at that specific temperature. This provides a quantified and accurate divisor for subsequent normalization calculations, ensuring the accuracy of temperature compensation.

[0042] More specifically, in a concrete example of this application, the process of obtaining the compensation factors is as follows: First, a pre-stored temperature compensation model function is retrieved from the device's non-volatile memory. This function was established in a laboratory environment through calibration tests on the same type of supercapacitor at different temperature points, and its form is a mathematical expression or lookup table. Then, the average temperature value calculated in the previous step is used as the independent variable and substituted into the temperature compensation model function. The compensation function calculations for resistance and capacitance are then performed separately, and the results of these calculations are the corresponding resistance compensation factor and capacitance compensation factor at the current temperature. These two calculated compensation factors will be directly used in the next step of parameter normalization calculation.

[0043] Accordingly, the original equivalent series resistance and original capacitance are divided by the resistance compensation factor and capacitance compensation factor, respectively, to obtain the normalized measurement vector. It should be understood that since the previous steps have obtained the original parameters affected by the current temperature and the compensation factor used to correct for this effect, but these two have not yet been combined to complete the removal of the temperature effect, the measured values ​​are still not standardized to a unified comparison benchmark. Therefore, in the technical solution of this application, the original equivalent series resistance and original capacitance are further divided by the resistance compensation factor and capacitance compensation factor, respectively, to obtain the normalized measurement vector, thereby performing a specific mathematical normalization operation to unify the discrete parameters measured under different temperature conditions to the benchmark of the standard reference temperature. In this way, a normalized measurement vector that eliminates the interference of temperature variables and has direct comparability can be finally generated, providing clean and standardized data input for subsequent filtering optimization and aging trend analysis on the time series.

[0044] More specifically, in a concrete example of this application, the process of generating the normalized measurement vector is implemented as follows. This process receives the original equivalent series resistance and resistance compensation factor calculated in the previous step. Then, a division operation is performed, using the value of the original equivalent series resistance as the dividend and the value of the resistance compensation factor as the divisor; the quotient is determined as the normalized equivalent series resistance. In parallel, the values ​​of the original capacitance and capacitance compensation factor are received. Another division operation is performed, using the original capacitance as the dividend and the capacitance compensation factor as the divisor; the quotient is determined as the normalized capacitance. Finally, the calculated values ​​of the normalized equivalent series resistance and the normalized capacitance are combined into a two-dimensional vector or data pair; this combination constitutes the normalized measurement vector and is input as an observation into the next step of the Kalman filter.

[0045] In step S420, a Kalman filter-based optimal state estimation is performed on the normalized measurement vector, the previous time-state estimate, and the previous time-covariance matrix to obtain the current time-state estimate. The current time-state estimate includes the current time-equivalent series resistance optimization estimate and the current time-capacitance optimization estimate. It should be understood that although the temperature-compensated normalized measurement vector eliminates temperature interference, as a single measurement result, it still inevitably contains random measurement noise introduced by sensor accuracy, signal acquisition, and parameter identification algorithms. Therefore, in this application's technical solution, a Kalman filter-based optimal state estimation is further performed on the normalized measurement vector, the previous time-state estimate, and the previous time-covariance matrix to obtain the current time-state estimate. This recursively fuses the current noisy measurement information with the optimal estimate of the system's historical state. This generates a statistically optimal current state estimate that effectively filters out random errors, thus enabling a smoother and more accurate tracking of the actual performance degradation trajectory of the supercapacitor due to aging.

[0046] More specifically, in a concrete example of this application, the state-optimal estimation based on a Kalman filter is decomposed into two consecutive stages: prediction and update. First, in the prediction stage, the equivalent series resistance and capacitance states at the current time are predicted a priori based on the state estimate from the previous time step and the state transition model describing the system's aging behavior. Simultaneously, the prior error covariance at the current time step is predicted by combining the error covariance matrix from the previous time step and the process noise covariance representing the uncertainty of the aging model. Then, in the update stage, the Kalman gain is first calculated, and its value determines the trust weight for the current normalized measurement vector. Next, using this Kalman gain, the residual between the predicted prior state value and the actual normalized measurement vector is weighted and corrected to calculate the optimal posterior state estimate for the current time step, which is the optimized estimate of the equivalent series resistance and the optimized estimate of the capacitance at the current time step. Finally, the error covariance matrix is ​​updated to reflect the reduction in system state uncertainty after the introduction of new measurement information; the updated matrix will be used in the prediction stage of the next cycle.

[0047] Specifically, in step S500, a health status assessment and early warning are performed based on the current state estimate to obtain the final health status and health alarm signal. It should be understood that a charging strategy using fixed parameters cannot adapt to the decline in the health status of the backup power supply itself, drastic changes in the operating environment temperature, and fluctuations in the grid state. Improper charging processes can accelerate its performance aging and even cause disturbances to the grid. Therefore, an intelligent control method capable of dynamically adjusting charging parameters based on real-time status is needed. Therefore, in the technical solution of this application, an adaptive charging strategy is further generated based on the final fused health status, the backup power supply surface temperature sequence, and the current grid state to obtain the charging command issued to the charger. This transforms accurate diagnostic results and real-time environmental perception into closed-loop refined control of the charging process, ensuring that the charging behavior always matches the current capacity of the backup power supply and external conditions. This enables proactive maintenance of the backup power supply, ensuring reliable charging while slowing its aging rate and maintaining friendly interaction with the grid, thereby extending its effective service life and improving the overall operational stability of the system.

[0048] Figure 5 This is a flowchart illustrating the method for monitoring the charging and discharging of the backup power supply in a wind turbine pitch system according to an embodiment of the present invention, which generates a buffer power command based on the junction temperature change rate and rotor angular velocity measurements. (See flowchart for example.) Figure 5 As shown, step S500 includes: S510, calculating the first SOH based on the current equivalent series resistance optimization estimate in the current state estimate; S520, calculating the second SOH based on the current capacitance optimization estimate in the current state estimate; S530, taking the minimum value of the first SOH and the second SOH as the final health state; S540, generating the health alarm signal based on the comparison between the final health state and a preset threshold.

[0049] In step S510, the first SOH is calculated based on the optimized estimate of the equivalent series resistance at the current time from the current time state estimate. This is expressed by the following formula:

[0050] in, The resistance threshold at which the lifespan ends. This is an optimized estimate of the equivalent series resistance at the current moment.

[0051] It should be understood that since the optimized estimate of the equivalent series resistance at the current moment obtained in the previous step is an absolute physical quantity, it cannot intuitively represent the health status of the supercapacitor relative to its new state and end-of-life state. Therefore, in the technical solution of this application, a first SOH is calculated based on the optimized estimate of the equivalent series resistance at the current moment in the current state estimate, thereby transforming the absolute resistance value into a standardized relative health status index expressed as a percentage. This calculation is achieved by dividing the difference between the resistance threshold at the end of life and the optimized estimate of the equivalent series resistance at the current moment by the difference between the resistance threshold at the end of life and the resistance value in the new state, thus obtaining the relative position of the current resistance state throughout its entire life cycle. In this way, the health status of the supercapacitor can be quantitatively assessed for the first time from the dimension of equivalent series resistance degradation, providing a key input basis for the final comprehensive health status determination.

[0052] In step S520, a second State of Health (SOH) is calculated based on the current-time capacitance optimization estimate from the current-time state estimate. It should be understood that, similar to the equivalent series resistance, the current-time capacitance optimization estimate obtained in the previous step is also an absolute physical quantity, and it cannot intuitively represent the health of the supercapacitor relative to its initial state and end-of-life standard in terms of energy storage capacity. Therefore, in the technical solution of this application, a second SOH is further calculated based on the current-time capacitance optimization estimate from the current-time state estimate, thereby converting this absolute energy storage capacity value into a standardized, percentage-based relative health status index. This allows for a second quantitative assessment of the supercapacitor's health status from the perspective of energy storage capacity degradation, providing another crucial input basis for the final comprehensive health status determination.

[0053] More specifically, in a concrete example of this application, the calculation process for the second SOH is implemented as follows. This calculation process first calls three key parameters: first, the current optimized capacitance estimate; second, the factory-set initial capacitance value of the supercapacitor preset in the system; and third, a preset capacitance threshold defining its lifespan end. At the start of the calculation, the current remaining effective capacitance range is obtained by subtracting the lifespan end capacitance threshold from the current optimized capacitance estimate. Then, the total effective capacitance range over the entire lifespan is obtained by subtracting the lifespan end capacitance threshold from the factory-set initial capacitance value. Finally, the calculated current remaining effective capacitance range is divided by the total effective capacitance range, and the quotient is multiplied by 100%. The result is the second SOH assessed from a capacitance perspective. This calculated second SOH value is used to compare with the first SOH to determine the final health status.

[0054] In steps S530 and S540, the minimum value of the first SOH and the second SOH is taken as the final health state; based on the comparison between the final health state and a preset threshold, the health alarm signal is generated. It should be understood that since the failure of a supercapacitor backup power supply may stem from an excessive increase in the equivalent series resistance or excessive decay of the capacitance, any performance dimension reaching its lower limit will lead to overall functional failure, and a numerical percentage of health status alone cannot directly translate into maintenance decisions. Therefore, in the technical solution of this application, the minimum value of the first SOH and the second SOH is further taken as the final health state; based on the comparison between the final health state and a preset threshold, the health alarm signal is generated, thereby following the barrel principle to make the most conservative assessment of the overall health status of the backup power supply, and converting this quantified health status into graded, operable maintenance instructions. This ensures that the final SOH assessment result does not overestimate the actual performance of the backup power supply and provides clear and timely early warning information to the maintenance system, thus achieving a closed loop from condition monitoring to predictive maintenance decisions.

[0055] More specifically, in a concrete example of this application, after calculating the first State of Health (SOH) representing the resistance health and the second State of Health (SOH) representing the capacitance health, the final health status determination and alarm signal generation process is performed as follows: First, the values ​​of the first and second SOHs are compared, and the smaller value is determined as the final health status. Then, the final health status value is sequentially compared with multi-level health status thresholds preset in the controller. For example, there are preset attention thresholds and replacement thresholds. If the final health status is higher than the attention threshold, a health alarm signal indicating a normal status is generated. If the final health status is lower than the attention threshold but higher than the replacement threshold, a health alarm signal is generated to remind maintenance personnel to pay attention during the next routine maintenance. If the final health status is lower than or equal to the replacement threshold, a high-priority health alarm signal requiring immediate replacement is generated. This finally generated alarm signal is sent to the wind turbine's main control or remote monitoring center to trigger the corresponding maintenance process.

[0056] In summary, the charging and discharging monitoring method for the backup power supply of a wind turbine pitch system according to an embodiment of the present invention is explained. When the wind turbine is in a preset safe state such as standby, a brief micro-pulse current is actively injected into the supercapacitor, thereby creating dynamic data for analysis without affecting the system's backup capability. By acquiring and analyzing the instantaneous voltage and current responses caused by this pulse at high frequency, two core indicators reflecting the health status of the supercapacitor—the original equivalent series resistance and the original capacitance—can be immediately identified. To eliminate the interference of ambient temperature on the measurement results, the method first normalizes the identified original parameters using a temperature compensation model to remove the temperature effect; then, a Kalman filter is used to optimize the normalized parameter sequence, filtering out random measurement noise and obtaining a smooth and reliable aging state estimate. Finally, based on the optimized resistance and capacitance states, the most accurate health status is comprehensively evaluated and an early warning is issued, achieving high-precision predictive diagnosis of early backup power supply failures.

[0057] Furthermore, a charging and discharging monitoring system for the backup power supply of a wind turbine pitch system is also provided.

[0058] Figure 6 This is a block diagram of a charging and discharging monitoring system for a backup power supply of a wind turbine pitch system according to an embodiment of the present invention. Figure 6 As shown, the charging and discharging monitoring system 100 for the backup power supply of a wind turbine pitch system according to an embodiment of the present invention includes: a test triggering and condition checking module 110, used to perform test triggering and condition checking based on the wind turbine state and state of charge to obtain a pulse injection command; a micro-pulse injection and high-frequency data acquisition module 120, used to perform micro-pulse injection and high-frequency data acquisition on the supercapacitor in response to the pulse injection command being true, to obtain an original dataset, wherein each data point in the original dataset includes the voltage across the supercapacitor, the current flowing through it, and the surface temperature; an original data acquisition module 130, used to perform instantaneous identification of key electrical parameters based on the original dataset to obtain the original equivalent series resistance, the original capacitance, and the average temperature; a parameter normalization and Kalman filter optimization module 140, used to perform parameter normalization and Kalman filter optimization on the original equivalent series resistance, the original capacitance, the average temperature, and the state estimate at the previous moment to obtain the state estimate at the current moment; and a health status assessment and early warning module 150, used to perform health status assessment and early warning based on the state estimate at the current moment to obtain the final health status and a health alarm signal.

[0059] As described above, the charging and discharging monitoring system 100 for the backup power supply of the wind turbine pitch system according to embodiments of the present invention can be deployed in an edge computing unit within the wind turbine hub, for example, in the controller of the pitch system itself or a dedicated monitoring unit, and can interact with the voltage and current sensors, surface temperature sensors, wind turbine main control system, and pitch charging module of the backup power supply in real time. In one possible implementation, the charging and discharging monitoring system 100 for the backup power supply of the wind turbine pitch system according to embodiments of the present invention can be integrated into the pitch control system of the wind turbine as an independent software module or hardware module. For example, the core models and parameters used for parameter identification and status assessment in this system, including the temperature compensation model function of the supercapacitor, the process and measurement noise covariance matrix of the Kalman filter, and the initial and lifespan termination thresholds for health status assessment, can be calibrated and set offline on the back-end server of the wind farm control center using component specifications and laboratory aging data. The optimized model parameter package can then be sent to the front-end monitoring unit. Similarly, the complete process for performing real-time online diagnostics in this system, including test triggering and condition checks, micro-pulse injection and high-frequency data acquisition, instantaneous identification of key electrical parameters, parameter normalization and Kalman filter optimization, and the final health status assessment and early warning signal generation, can also be embedded in dedicated edge computing hardware, such as a digital signal processor or field-programmable gate array module inside the pitch controller. This accelerates the calculation process from data acquisition to parameter identification and filter optimization, ensuring low-latency generation and output of the final health alarm signal.

[0060] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of the present invention, and the present invention is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A method for monitoring the charging and discharging of a backup power supply for a wind turbine pitch system, characterized in that, include: Test triggering and condition checking are performed based on the wind turbine status and charge status to obtain pulse injection commands; In response to the pulse injection command being true, the charging module performs micro-pulse injection and high-frequency data acquisition on the supercapacitor to obtain a raw dataset. Each data point in the raw dataset includes the voltage across the supercapacitor, the current flowing through it, and the surface temperature. Instantaneous identification of key electrical parameters is performed based on the original dataset to obtain the original equivalent series resistance, original capacitance, and average temperature. The original equivalent series resistance, original capacitance, average temperature and the state estimate of the previous time step are normalized and optimized by Kalman filtering to obtain the state estimate of the current time step. Health status assessment and early warning are performed based on the current state estimate to obtain the final health status and health alarm signal.

2. The method for monitoring the charging and discharging of the backup power supply for a wind turbine pitch system according to claim 1, characterized in that, Test triggering and condition checks are performed based on the wind turbine status and state of charge to obtain pulse injection commands, including: Determine whether the fan status belongs to the set of fan statuses that are allowed to be tested to obtain the first judgment result; Determine whether the state of charge exceeds the minimum state of charge threshold to obtain a second determination result; In response to both the first and second judgment results being true, a system ready flag is generated; Perform compliance calculation and judgment on the test cycle interval between the current timestamp and the last test timestamp to obtain the interval ready mark; The pulse injection command is generated based on the interval readiness marker and the current timestamp.

3. The method for monitoring the charging and discharging of the backup power supply for a wind turbine pitch system according to claim 1, characterized in that, Instantaneous identification of key electrical parameters based on the original dataset is performed to obtain the original equivalent series resistance, original capacitance, and average temperature, including: Pulse event features were extracted from the original dataset to obtain the pulse start timestamp, pulse end timestamp, pulse current amplitude, instantaneous voltage before the pulse, instantaneous voltage at the beginning of the pulse, instantaneous voltage at the end of the pulse, and temperature sequence during the pulse. Calculate the original equivalent series resistance based on the pulse current amplitude, the instantaneous voltage before the pulse, and the instantaneous voltage at the beginning of the pulse; Calculate the original capacitance based on the pulse start timestamp, pulse end timestamp, pulse current amplitude, pulse initial instantaneous voltage, and pulse final instantaneous voltage; The average temperature is calculated based on the temperature sequence during the pulse.

4. The method for monitoring the charging and discharging of the backup power supply for a wind turbine pitch system according to claim 3, characterized in that, The original equivalent series resistance is calculated based on the pulse current amplitude, the instantaneous voltage before the pulse, and the instantaneous voltage at the beginning of the pulse. This includes calculating the original equivalent series resistance using the following formula: ， ， in, The instantaneous voltage before the pulse. The voltage at the initial instant of the pulse. The amplitude of the pulse current. The initial pressure difference of the pulse. This is the original equivalent series resistance.

5. The method for monitoring the charging and discharging of the backup power supply for a wind turbine pitch system according to claim 3, characterized in that, The original capacitance is calculated based on the pulse start timestamp, pulse end timestamp, pulse current amplitude, initial pulse voltage, and final pulse voltage. This includes calculating the original capacitance using the following formula: ， ， in, and These are the pulse end timestamp and the pulse start timestamp, respectively. The amplitude of the pulse current. The voltage at the initial instant of the pulse. The voltage at the end of the pulse. The instantaneous pressure difference of the pulse. This is the original capacitor.

6. The method for monitoring the charging and discharging of the backup power supply for a wind turbine pitch system according to claim 1, characterized in that, The original equivalent series resistance, original capacitance, average temperature, and the state estimate from the previous time step are normalized and optimized using Kalman filtering to obtain the current state estimate, including: The parameters of the original equivalent series resistance, original capacitance, and average temperature are normalized based on a temperature compensation model to obtain a normalized measurement vector. The state-optimal estimation based on the Kalman filter is performed on the normalized measurement vector, the state estimate of the previous time step, and the covariance matrix of the previous time step to obtain the state estimate of the current time step. The state estimate of the current time step includes the optimized estimate of the equivalent series resistance and the optimized estimate of the capacitance at the current time step.

7. The method for monitoring the charging and discharging of the backup power supply for a wind turbine pitch system according to claim 6, characterized in that, The original equivalent series resistance, original capacitance, and average temperature are normalized using a temperature compensation model to obtain a normalized measurement vector, including: The average temperature is input into the temperature compensation model function of the supercapacitor to obtain the resistance compensation factor and the capacitance compensation factor. The normalized measurement vector is obtained by dividing the original equivalent series resistance and the original capacitance by the resistance compensation factor and the capacitance compensation factor, respectively.

8. The method for monitoring the charging and discharging of the backup power supply for a wind turbine pitch system according to claim 1, characterized in that, Based on the current state estimate, a health status assessment and early warning are performed to obtain the final health status and health alarm signal, including: The first State of Health (SOH) is calculated based on the optimized estimate of the equivalent series resistance at the current time from the current state estimate. Calculate the second SOH based on the current capacitance optimization estimate from the current state estimate; The minimum value of the first SOH and the second SOH is taken as the final health state; The health alarm signal is generated based on the comparison between the final health status and the preset threshold.

9. The method for monitoring the charging and discharging of the backup power supply for a wind turbine pitch system according to claim 8, characterized in that, The first State of Health (SOH) is calculated based on the optimized estimate of the equivalent series resistance at the current moment from the current state estimate. This includes: calculating the first SOH using the optimized estimate of the equivalent series resistance at the current moment from the current state estimate using the following formula: ， in, This is the resistance threshold at the end of the lifespan. The optimal estimate of the equivalent series resistance at the current moment. To obtain the maximum value, It is the first SOH.

10. A charging and discharging monitoring system for a backup power supply of a wind turbine pitch system, characterized in that, include: The test triggering and condition checking module is used to perform test triggering and condition checking based on the wind turbine status and charge status to obtain pulse injection commands; The micro-pulse injection and high-frequency data acquisition module is used to respond to the pulse injection command being true, and the charging module performs micro-pulse injection and high-frequency data acquisition on the supercapacitor to obtain the original dataset. Each data point in the original dataset includes the voltage across the supercapacitor, the current flowing through it, and the surface temperature. The raw data acquisition module is used to perform instantaneous identification of key electrical parameters based on the raw dataset to obtain the raw equivalent series resistance, raw capacitance, and average temperature. The parameter normalization and Kalman filter optimization module is used to perform parameter normalization and Kalman filter optimization on the original equivalent series resistance, original capacitance, average temperature and the state estimate of the previous time step to obtain the state estimate of the current time step. The health status assessment and early warning module is used to assess and warn of health status based on the current status estimate to obtain the final health status and health alarm signal.

Images