Wind turbine generator converter switching tube open-circuit fault detection method and system

Abstract

The invention relates to the field of converter fault detection, and provides a wind turbine generator converter switching tube open-circuit fault detection method and system, and the method comprises the steps: comparing the actual output current of a converter with the internal instruction current to judge whether the converter is out of control or not, thereby precisely distinguishing the internal fault from the external disturbance; the method comprises the following steps: firstly, calculating a residual error between an actual current and a reference current on a d-q axis, wherein the residual error directly reflects the degree that an actual behavior deviates from a system intention; in order to eliminate the influence of wind speed and load change on the absolute value of the residual error, normalization processing is carried out on the residual error by using reference current magnitude, and a normalized fault characteristic quantity with consistency under all working conditions is constructed; in order to improve the detection accuracy and avoid misjudgment caused by instantaneous disturbance, sliding window integration is carried out on the normalized fault characteristic quantity, and the occurrence of a fault is confirmed only when an integral value continuously exceeds a preset threshold value; once the fault is confirmed, the phase angle information of the current residual vector is combined with the voltage phase angle of the power grid to quickly and accurately position a specific fault switch tube.

Description

Technical Field

[0001] This invention relates to the field of converter fault detection technology, and in particular to a method and system for detecting open-circuit faults in the switching tubes of wind turbine converters. Background Technology

[0002] Wind power, as a clean and renewable energy source, occupies an increasingly important position in the global energy structure. The converter in a wind turbine is the core component for achieving efficient conversion of wind energy into electrical energy and ensuring power quality. The converter consists of multiple power electronic switching transistors (such as IGBTs). These transistors switch frequently under harsh operating conditions, making them a weak link in the system, prone to open-circuit or short-circuit faults. An open-circuit fault in a single transistor can disrupt the three-phase symmetry of the converter's output current, causing current distortion, torque pulsation, and DC bus voltage fluctuations. In severe cases, it can trigger overcurrent or overvoltage protection, leading to unplanned shutdowns of the wind turbine, reduced power generation efficiency, and even a chain reaction that damages other critical components within the converter, causing significant economic losses. Therefore, developing a fast and accurate method for detecting open-circuit faults in wind turbine converter switching transistors is crucial for ensuring the safe and stable operation of wind turbines.

[0003] Various converter fault detection schemes have been proposed in the existing technology, most of which are based on the analysis of the converter output current signal. These methods extract fault characteristics by monitoring the distortion of the current waveform, calculating the average value or symmetrical component of the current, or analyzing the trajectory change of the Park vector in the dq coordinate system. However, these traditional methods face a serious technical challenge in practical applications: it is difficult to effectively distinguish between intrinsic current anomalies caused by open-circuit faults of the switching transistors and extrinsic current anomalies caused by external disturbances such as grid voltage dips and frequency fluctuations. Under the requirements of modern wind power grid connection standards (such as low voltage ride-through), when the grid is disturbed, the converter control system will actively adjust its output to provide dynamic support to the grid. This process itself will generate non-sinusoidal and asymmetrical currents. However, traditional converter fault detection methods only observe the final output current and cannot distinguish whether this current distortion is a normal behavior actively executed by the control system in response to external disturbances or an unexpected result caused by internal component failure. This can easily lead to false alarms, causing wind turbines to disconnect from the grid when they should not, affecting grid stability and unit availability. Summary of the Invention

[0004] The present invention aims to solve at least one of the problems existing in the prior art, and provides a method and system for detecting open circuit faults in the switching tubes of wind turbine converters.

[0005] One aspect of the present invention provides a method for detecting open-circuit faults in the switching transistors of a wind turbine converter, the method comprising: The actual three-phase current output from the wind turbine converter is transformed by coordinates to obtain the actual current along the dq axis. Calculate the dq-axis current residual between the actual dq-axis current and the dq-axis reference current; Based on the dq-axis reference current, the residual vector of the dq-axis current is normalized and its characteristics are constructed to obtain normalized fault characteristic quantities. The normalized fault feature quantity is integrated with the fault feature quantity and a threshold is determined to obtain the fault identifier; the fault identifier is used to indicate whether a fault has been detected. In response to the fault flag being true, fault location is determined based on the dq-axis current residual and the phase angle of the grid voltage vector.

[0006] Optionally, a coordinate transformation is performed on the actual three-phase current output by the wind turbine converter to obtain the actual dq-axis current, including: Extract the grid voltage vector phase angle from the phase-locked loop module of the wind turbine; Based on the phase angle of the grid voltage vector, the actual three-phase currents are transformed using the following formula to obtain the actual dq-axis currents: ; in, The phase angle of the grid voltage vector. These are the actual currents of phases A, B, and C, respectively. These are the actual currents along the d-axis and q-axis, respectively.

[0007] Optionally, based on the dq-axis reference current, residual vector normalization and feature construction are performed on the dq-axis current residual to obtain normalized fault characteristic quantities, including: Calculate the reference current for the dq axis reference current; Calculate the magnitude of the current residual vector of the dq-axis current residual; The normalized fault characteristic quantity is obtained by dividing the magnitude of the current residual vector by the reference current quantity.

[0008] Optionally, the reference current quantity for calculating the dq-axis reference current includes: The reference current for the dq axis is calculated using the following formula: ; in, and These are the d-axis reference current and the q-axis reference current, respectively. This is the reference current.

[0009] Optionally, fault feature integration and threshold judgment are performed on the normalized fault feature quantities to obtain the fault identifier, including: Set a sliding time window; Within the sliding time window, the normalized fault feature quantity is integrated to obtain the integrated fault feature. The integrated fault characteristics are compared with a fixed fault threshold to obtain the fault identifier.

[0010] Optionally, within the sliding time window, the normalized fault feature quantity is integrated to obtain the integrated fault features, including: The fault characteristic integral is performed on the normalized fault characteristic quantity using the following formula: ; in, To normalize fault characteristic quantities, For integration variables, For sliding time windows, for Integral fault characteristics at any given time.

[0011] Optionally, the integral fault characteristics are compared with a fixed fault threshold to obtain a fault identifier, including: In response to the duration for which the integral fault characteristic exceeds a fixed fault threshold for a predetermined period of time, the fault flag is set to true.

[0012] Optionally, in response to the fault flag being true, fault location is determined based on the dq-axis current residual and the grid voltage vector phase angle, including: Calculate the phase angle of the residual vector of the dq-axis current residual in the dq coordinate system; The phase angle of the residual vector is subjected to inverse Park transformation to obtain the electrical angle value; The electrical angle value is mapped to a specific converter switch to determine the fault location.

[0013] Optionally, mapping the electrical angle range to a specific converter switch to obtain the fault location includes: According to the preset mapping table, the specific converter switch corresponding to the electrical angle value is taken as the fault location; The preset mapping table stores the unique correspondence between six consecutive 60-degree sectors in a 360-degree electrical cycle and six switching transistors.

[0014] Another aspect of the present invention provides a wind turbine converter switch open circuit fault detection system, the wind turbine converter switch open circuit fault detection system comprising: The coordinate transformation module is used to perform coordinate transformation on the actual three-phase current output by the wind turbine converter to obtain the actual current of the dq axis. The current residual calculation module is used to calculate the dq-axis current residual between the actual dq-axis current and the dq-axis reference current. The vector normalization module is used to perform residual vector normalization and feature construction on the dq axis current residual based on the dq axis reference current, so as to obtain normalized fault feature quantities. The threshold determination module is used to perform fault feature integration and threshold determination on the normalized fault feature quantity to obtain the fault identifier; the fault identifier is used to indicate whether a fault has been detected. The fault location module is used to locate the fault position based on the dq-axis current residual and the phase angle of the grid voltage vector in response to the fault identifier being true.

[0015] Compared with existing technologies, this invention provides a method and system for detecting open-circuit faults in wind turbine converter switching transistors. It determines whether the circuit is out of control by comparing the actual output current with the internal command current of the control system, thereby accurately distinguishing between internal faults and external disturbances. First, the residual between the actual current and the reference current on the dq axis is calculated. This residual directly reflects the degree to which the actual behavior deviates from the system's intended behavior. To eliminate the influence of wind speed and load changes on the absolute value of the residual, the residual vector is further normalized using the reference current, constructing a normalized fault characteristic quantity that is consistent across all operating conditions. To improve detection accuracy and avoid misjudgments caused by instantaneous disturbances, a sliding window integration is performed on this normalized fault characteristic quantity. Only when the integral value continuously exceeds a preset threshold is the occurrence of a fault confirmed. Once the fault is confirmed, the specific faulty switching transistor can be quickly and accurately located using the phase angle information of the current residual vector combined with the grid voltage phase angle. Attached Figure Description

[0016] One or more embodiments are illustrated by way of example with the corresponding pictures in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0017] Figure 1 A flowchart of a method for detecting open-circuit faults in the switching transistors of a wind turbine generator according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the data flow in the wind turbine converter switch tube open-circuit fault detection method according to an embodiment of the present invention; Figure 3The flowchart describes a method for detecting open-circuit faults in the converter switching tubes of a wind turbine according to an embodiment of the present invention. Based on the dq-axis reference current, the residual vector of the dq-axis current is normalized and its features are constructed to obtain the normalized fault feature quantity. Figure 4 This is a flowchart illustrating the method for detecting open-circuit faults in the converter switching tubes of a wind turbine according to an embodiment of the present invention, which involves performing fault feature integration and threshold judgment on normalized fault feature quantities to obtain fault identifiers. Figure 5 This is a block diagram of a wind turbine converter switch tube open-circuit fault detection system according to an embodiment of the present invention. Detailed Implementation

[0018] 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.

[0019] As indicated in the specification and claims of this invention, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0020] While this invention makes various references to certain modules in systems according to embodiments of the invention, any number of different modules can be used and run on user terminals and / or servers. The modules are merely illustrative, and different aspects of the systems and methods may use different modules.

[0021] This invention uses flowcharts to illustrate the operations performed by the system according to embodiments of the invention. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously, as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0022] In response to the technical problems in the background art, the present invention proposes a wind turbine converter switch tube open circuit fault detection scheme to solve the technical problem in the prior art that it is difficult to effectively distinguish between internal faults such as internal current distortion caused by the open circuit fault of the converter switch tube itself and external disturbances caused by grid voltage sags, frequency fluctuations, etc. Specifically, the open-circuit fault detection scheme for wind turbine converter switching transistors proposed in this invention first starts from the data source, delving into the internal control system of the wind turbine to acquire the current reference command and actual current of the dq axis in real time, and calculates the current residual between the two. This current residual accurately quantifies the deviation between the physical execution of the converter and the intention of the controller. Considering the variable operating conditions of wind turbines, to ensure the consistency of detection standards, this invention further normalizes the current residual by using the modulus of the reference current, thereby constructing a standardized normalized fault characteristic quantity that is unaffected by the load size. To improve the reliability of diagnosis and filter out noise interference, a sliding time window is introduced to integrate the normalized characteristic quantity, ensuring that the integral value accumulates and triggers the fault indicator only when the fault characteristic persists. Once the fault is confirmed, the phase angle information of the current residual is immediately used, combined with the phase angle of the grid voltage vector obtained from the phase-locked loop, and the fault characteristic is accurately mapped to the corresponding electrical angle range through the inverse coordinate transformation, ultimately achieving rapid location of the specific faulty switching transistor.

[0023] The present invention proposes a method for detecting open-circuit faults in the switching tubes of wind turbine converters. Figure 1 This is a flowchart of a method for detecting open-circuit faults in the switching transistors of a wind turbine converter according to an embodiment of the present invention. Figure 2 This is a schematic diagram of data flow in a wind turbine converter switch open-circuit fault detection method according to an embodiment of the present invention. (In conjunction with...) Figure 1 and Figure 2 According to an embodiment of the present invention, a method for detecting open-circuit faults in the switching transistors of a wind turbine converter includes the following steps: S100, performing coordinate transformation on the actual three-phase current output by the wind turbine converter to obtain the actual dq-axis current; S200, calculating the dq-axis current residual between the actual dq-axis current and the dq-axis reference current; S300, based on the dq-axis reference current, performing residual vector normalization and feature construction on the dq-axis current residual to obtain a normalized fault feature quantity; S400, performing fault feature integration and threshold judgment on the normalized fault feature quantity to obtain a fault identifier; the fault identifier is used to indicate whether a fault is detected; S500, in response to the fault identifier being true, performing fault location based on the dq-axis current residual and the phase angle of the grid voltage vector to obtain the fault location.

[0024] Specifically, in step S100, the actual three-phase current output by the wind turbine converter is transformed using coordinates to obtain the actual current along the dq axis. It should be understood that the core control strategy of the wind turbine converter, such as the decoupling control of active and reactive power, is completed in the dq synchronous rotating coordinate system. The generated current reference command is itself a DC component along the dq axis, while the actual three-phase current output by the wind turbine converter, directly acquired by the sensor, is a time-varying AC sine wave. This cannot be directly and meaningfully compared with the current reference command. Therefore, in the technical solution of this invention, the actual three-phase current output by the wind turbine converter is transformed using coordinates to obtain the actual current along the dq axis, thereby unifying the expression of the actual current and the command form of the reference current under the same coordinate system. This allows the complex AC current signal to be converted into a DC component that is easy to analyze and compare under steady-state conditions, laying the foundation for subsequent accurate calculation of the residual between the actual current and the reference command, and enabling any abnormal behavior deviating from the control intention to be clearly quantified.

[0025] More specifically, in this embodiment of the invention, the coordinate transformation of the three-phase actual current output by the wind turbine converter to obtain the dq-axis actual current includes: extracting the grid voltage vector phase angle from the phase-locked loop module of the wind turbine; and performing coordinate transformation on the three-phase actual current based on the grid voltage vector phase angle to obtain the dq-axis actual current.

[0026] Specifically, the phase angle of the grid voltage vector is extracted from the phase-locked loop (PLL) module of the wind turbine. It should be understood that the subsequent coordinate transformation (i.e., the Park transformation) is a mathematical process that converts AC quantities in a stationary coordinate system to DC quantities in a synchronously rotating coordinate system. The accuracy of this mathematical process depends entirely on a rotation angle that can accurately represent the instantaneous position of the grid voltage vector. If this rotation angle has a deviation or delay, the transformed dq-axis current will not accurately reflect the true state of the wind turbine system, introducing unwanted pulsating components, thus severely interfering with subsequent residual calculations and fault feature extraction. Therefore, in the technical solution of this invention, the phase angle of the grid voltage vector is further extracted from the PLL module of the wind turbine to provide a high-precision reference angle that is synchronized with the grid in real time for the coordinate transformation. This ensures that the actual three-phase currents can be accurately projected onto the dq coordinate system that is perfectly aligned with the control commands, providing a guarantee for subsequent unbiased command-response consistency comparison.

[0027] More specifically, in a specific example of the present invention, when extracting the grid voltage vector phase angle from the phase-locked loop module of the wind turbine, firstly, the three-phase voltage signal on the grid side is acquired and converted to a two-phase stationary state using Clarke transformation. coordinate system, to obtain The voltage components in the coordinate system; then, using the phase angle of the grid voltage vector output by the phase-locked loop at the previous moment, the voltage components are... The voltage components in the coordinate system are subjected to Park transformation to obtain the voltage vector in the dq rotating coordinate system, including the d-axis voltage components. and q-axis voltage component The goal of phase-locked loop control is to align the dq coordinate system with the grid voltage vector. At this point, the q-axis voltage component of the voltage vector... It should be zero. Therefore, As the input to the PI regulator, closed-loop control is performed through the PI regulator, and the output of the PI regulator is the estimated angular frequency of the power grid. Finally, by integrating the estimated angular frequency, the real-time updated phase angle of the power grid voltage vector can be obtained, which is then provided to the current coordinate transformation process.

[0028] Specifically, based on the phase angle of the grid voltage vector, a coordinate transformation is performed on the actual three-phase current to obtain the actual dq-axis current, including: based on the phase angle of the grid voltage vector, a coordinate transformation is performed using the following formula to obtain the actual dq-axis current: ; in, The phase angle of the grid voltage vector. These are the actual currents of phases A, B, and C, respectively. These are the actual currents along the d-axis and q-axis, respectively.

[0029] It should be understood that the actual three-phase current directly measured by the sensor is a time-varying AC sine wave in a stationary coordinate system, while the reference current command used for decoupling control within the converter control system is a DC quantity in a synchronous rotating coordinate system (dq). These two types of data are different in both physical dimension and mathematical form, and cannot be directly compared to determine whether the system is operating normally. Therefore, in the technical solution of this invention, the actual three-phase current is further transformed based on the phase angle of the grid voltage vector to obtain the actual current along the dq axis. This transforms the measured physical quantity into a mathematical framework completely identical to the control command. In this way, the complex AC current waveform can be converted into a DC component that is essentially constant in steady state, allowing the dynamic response of the actual current to be directly and in real-time algebraically compared with the DC reference command. This facilitates the subsequent accurate calculation of the current residual, which can quantify the system execution deviation.

[0030] Specifically, in step S200, the residual current between the actual dq-axis current and the reference dq-axis current is calculated. It should be understood that simply observing the absolute value or fluctuations of the actual dq-axis current is insufficient to fundamentally determine whether the dynamic behavior of the current is caused by commands actively issued by the converter control system in response to external grid disturbances, or by an uncontrolled result due to an open-circuit fault in the converter's internal switching transistors. This introduces inherent ambiguity in fault diagnosis. Therefore, in the technical solution of this invention, the residual current between the actual dq-axis current and the reference dq-axis current is further calculated to directly quantify the degree of deviation between the converter's physical execution level and the controller's intended commands. This generates a residual signal that remains small when the converter control system is operating normally or responding to external disturbances (when the actual current effectively tracks the reference current), but increases when an internal open-circuit fault occurs in the converter (when physical limitations prevent the actual current from tracking the reference current). This provides a clear and unambiguous physical criterion for distinguishing between internal converter faults and external disturbances.

[0031] More specifically, in a concrete example of the present invention, firstly, within the current control cycle, the arithmetic unit of the digital signal processor (DSP) reads two key data points from its internal memory: one is the actual dq-axis current value just obtained through coordinate transformation, and the other is the dq-axis reference current value generated by the upper-level power control loop for the current cycle. Next, the arithmetic unit performs parallel subtraction operations, subtracting the actual d-axis current value from the d-axis reference current value to obtain the d-axis current residual; simultaneously, it subtracts the actual q-axis current value from the q-axis reference current value to obtain the q-axis current residual. Finally, the calculated d-axis current residual and q-axis current residual form a two-dimensional vector, which is then passed to the subsequent residual vector normalization and feature construction process for processing.

[0032] Specifically, in step S300, based on the dq-axis reference current, the dq-axis current residual is subjected to residual vector normalization and feature construction to obtain a normalized fault characteristic quantity. It should be understood that the absolute amplitude of the dq-axis current residual is closely related to the real-time operating conditions (i.e., output power) of the wind turbine. A fixed current residual is within the normal fluctuation range under heavy load, but signifies a serious fault under light load. This makes it difficult to set a fixed fault threshold that can adapt to all operating conditions if the current residual is directly used as the criterion. Therefore, in the technical solution of this invention, the dq-axis current residual is further subjected to residual vector normalization and feature construction based on the dq-axis reference current to obtain a normalized fault characteristic quantity. This eliminates the influence of load changes on the fault characteristic amplitude, transforming the absolute deviation into a relative, standardized deviation rate. This allows for the generation of a normalized fault characteristic quantity with consistent scaling and high sensitivity across the entire operating range of the wind turbine, from light load to full load. This enables reliable fault judgment using a fixed fault threshold, improving the operational adaptability and accuracy of the detection algorithm.

[0033] Figure 3 This document describes a flowchart illustrating the process of normalizing and constructing the residual vector of the dq-axis current residual based on the dq-axis reference current in a wind turbine converter switch open-circuit fault detection method according to an embodiment of the present invention, to obtain the normalized fault characteristic quantity. For example... Figure 3 As shown, step S300 includes: S310, calculating the reference current of the dq-axis reference current; S320, calculating the magnitude of the current residual vector of the dq-axis current residual; S330, dividing the magnitude of the current residual vector by the reference current to obtain the normalized fault characteristic quantity.

[0034] Specifically, in step S310, the reference current of the dq-axis reference current is calculated, including: calculating the reference current of the dq-axis reference current using the following formula: ; in, and These are the d-axis reference current and the q-axis reference current, respectively. This is the reference current.

[0035] It should be understood that subsequent normalization requires a scalar that can accurately characterize the overall amplitude of the current converter command current as a reference. However, the dq-axis reference current exists as two mutually orthogonal components and cannot be directly used as a normalization factor. Therefore, in the technical solution of this invention, the reference current quantity of the dq-axis reference current is further calculated to combine the two orthogonal current command components of the d-axis and q-axis into a scalar value that can comprehensively reflect the overall intensity of the current command current, i.e., the reference current quantity. In this way, an accurate and dynamically changing reference quantity can be provided for the subsequent residual vector normalization step. This reference quantity is directly related to the real-time operating point of the wind turbine, ensuring that the finally constructed normalized fault characteristic quantity can effectively escape the influence of changes in operating conditions.

[0036] Specifically, in step S320, the magnitude of the current residual vector of the dq-axis current residual is calculated. It should be understood that the dq-axis current residual exists as two mutually orthogonal components on the d and q axes. Analyzing either component alone cannot comprehensively measure the overall severity of the deviation of the actual current vector from the reference current vector, and it is difficult to make a unified threshold judgment. Therefore, in the technical solution of this invention, the magnitude of the current residual vector of the dq-axis current residual is further calculated to synthesize the two-dimensional residual vector information into a non-directional scalar value. This scalar value uniquely represents the Euclidean distance between the actual current and the command current at the current moment. In this way, a single feature quantity that can intuitively and comprehensively reflect the severity of the fault can be generated, providing a simple and effective input for subsequent normalization processing and threshold comparison.

[0037] More specifically, in a specific example of the present invention, firstly, the DSP retrieves the values ​​of the d-axis current residual and the q-axis current residual obtained in the previous calculation step from its internal register; secondly, the DSP performs two multiplication operations to obtain the square values ​​of the d-axis current residual and the q-axis current residual, respectively; subsequently, it performs an addition operation to add the two square values ​​to obtain a sum of squares; finally, it calls a square root function or uses hardware acceleration instructions to perform a square root operation on the sum of squares, and the final calculation result is the current residual vector magnitude that characterizes the overall size of the current residual. This scalar value is stored and used for subsequent normalization calculations.

[0038] Specifically, in step S330, the magnitude of the current residual vector is divided by the reference current to obtain the normalized fault characteristic quantity. It should be understood that the absolute value of the unprocessed magnitude of the current residual vector is strongly correlated with the real-time operating conditions of the wind turbine (e.g., output power), leading to significant differences in the characteristic amplitude of the same fault under different loads, making fault judgment using a fixed threshold unreliable. Therefore, in the technical solution of this invention, the magnitude of the current residual vector is further divided by the reference current to obtain the normalized fault characteristic quantity, thereby converting the absolute residual magnitude into a relative deviation rate relative to the current command current amplitude. This generates a normalized fault characteristic quantity with a consistent scale under different operating conditions. This normalized fault characteristic quantity can stably reflect the relative severity of the fault, thus providing a solid foundation for reliable judgment using a fixed threshold and improving the accuracy and adaptability of the diagnostic method across all operating conditions.

[0039] Specifically, in step S400, the normalized fault characteristic quantity is integrated and threshold-judged to obtain a fault identifier, which indicates whether a fault has been detected. It should be understood that the normalized fault characteristic quantity is still subject to interference from sensor noise, transient disturbances in the control loop, or high-frequency harmonics during actual operation, manifesting as brief, non-continuous spike pulses. If judgment is based solely on instantaneous values, misjudgments can occur, leading to unnecessary shutdown protection actions. Therefore, in the technical solution of this invention, the normalized fault characteristic quantity is further integrated and threshold-judged to obtain a fault identifier. This utilizes the time accumulation effect to smooth and suppress these transient interferences, ensuring that a fault is only confirmed as a real fault when the fault characteristic persists and accumulates to a certain extent. This enhances the accuracy and reliability of fault diagnosis, effectively avoids false alarms caused by accidental signal glitches, ensures that the final output fault identifier has high confidence, and provides a solid and reliable decision-making basis for subsequent precise fault location and protection strategies.

[0040] Figure 4 This is a flowchart illustrating the method for detecting open-circuit faults in wind turbine converter switching transistors according to an embodiment of the present invention, which involves integrating normalized fault characteristic quantities and determining a threshold to obtain a fault identifier. Figure 4 As shown, step S400 includes: S410, setting a sliding time window; S420, performing fault feature integration on the normalized fault feature quantity within the sliding time window to obtain integrated fault features; S430, comparing the integrated fault features with a fixed fault threshold to obtain a fault identifier.

[0041] Specifically, in steps S410 and S420, a sliding time window is set; within the sliding time window, the normalized fault feature quantity is integrated to obtain the integrated fault feature. Specifically, integrating the normalized fault feature quantity within the sliding time window to obtain the integrated fault feature includes: integrating the normalized fault feature quantity using the following formula: ; in, To normalize fault characteristic quantities, For integration variables, For sliding time windows, for Integral fault characteristics at any given time.

[0042] It is understandable that transient electrical noise or minor disturbances within the converter control system can cause brief, non-continuous spikes or glitches in the normalized fault characteristic quantity generated in the previous step. Directly comparing these instantaneous values ​​with the threshold can trigger erroneous fault judgments due to occasional signal fluctuations. Therefore, in the technical solution of this invention, a sliding time window is further set; within the sliding time window, the normalized fault characteristic quantity is integrated to obtain the integral fault characteristic. This utilizes the cumulative effect over time to smooth out the influence of random noise and transient interference, while amplifying the persistent true fault characteristics. This generates an integral fault characteristic with a higher signal-to-noise ratio that more stably reflects the fault state of the converter control system, providing a more accurate decision-making basis for subsequent reliable threshold judgments, thereby effectively avoiding false alarms caused by transient interference.

[0043] Specifically, in step S430, the integral fault characteristics are compared with a fixed fault threshold to obtain a fault identifier. It should be understood that even after integral smoothing, certain severe but non-faulty transient system disturbances can still cause the integral fault characteristics to briefly exceed the threshold. If a judgment is made based on a single comparison, there is still a risk of false alarms. Therefore, in the technical solution of this invention, the integral fault characteristics are further compared with a fixed fault threshold to obtain a fault identifier. This adds a time-dimensional confirmation step to the amplitude judgment, ensuring that a fault decision is made only when the severity of the fault characteristic not only reaches the threshold but also its duration meets the preset minimum fault duration requirement. This provides a final line of defense for fault diagnosis, filtering out all non-persistent interference events to the greatest extent possible, ensuring that the final generated fault identifier has high confidence, and thus providing an absolutely reliable trigger signal for the converter control system to initiate subsequent fault location or protection strategies, avoiding unnecessary shutdowns due to misjudgments.

[0044] More specifically, in this embodiment of the invention, comparing an integral fault feature with a fixed fault threshold to obtain a fault identifier includes: setting the fault identifier to true in response to the duration for which the integral fault feature exceeds the fixed fault threshold for a predetermined period of time.

[0045] Specifically, in response to the duration for which the integral fault characteristic exceeds a fixed fault threshold for a predetermined period, the fault flag is set to true. It should be understood that in certain extreme but brief system transient processes, such as severe grid voltage distortion or sudden load surges, the integral fault characteristic may still briefly exceed the preset fixed fault threshold. If the fault flag is triggered based solely on a single over-limit event, there is still a risk of misjudgment due to non-persistent disturbances. Therefore, in the technical solution of this invention, the fault flag is further set to true in response to the duration for which the integral fault characteristic exceeds the fixed fault threshold for a predetermined period. This adds a time-dimensional confirmation logic to the fault judgment, ensuring that only those abnormal states that not only meet the fault criteria in amplitude but also exhibit persistence in time are responded to. This effectively prevents false alarms and filters out all non-persistent transient events, ensuring that the final output fault flag has high reliability and confidence, thereby providing an accurate decision-making basis for subsequently initiating critical protection and location procedures.

[0046] Specifically, in step S500, in response to a true fault flag, fault location is determined based on the dq-axis current residual and the phase angle of the grid voltage vector. It should be understood that upon confirming a fault occurrence (i.e., when the fault flag is true), only a Boolean flag indicating an anomaly is obtained, lacking diagnostic information about the specific fault location. This is insufficient for implementing subsequent targeted fault-tolerant control strategies or guiding maintenance personnel for rapid repair. Therefore, in the technical solution of this invention, in response to a true fault flag, fault location is further determined based on the dq-axis current residual and the phase angle of the grid voltage vector. This allows for deeper analysis of the directional information inherent in the residual vector itself, transforming this directional information from the abstract dq-axis rotating coordinate system back to the three-phase stationary coordinate system directly related to the physical switch position. This establishes a unique mapping between the electrical distribution of fault characteristics and the specific switch, enabling a leap from general fault detection to precise fault location. This provides crucial and actionable diagnostic conclusions for implementing advanced fault-tolerant measures such as isolating faulty bridge arms, adjusting control algorithms, or generating accurate maintenance work orders.

[0047] More specifically, in this embodiment of the invention, in response to a true fault flag, fault location is determined based on the dq-axis current residual and the grid voltage vector phase angle to obtain the fault location, including: calculating the residual vector phase angle of the dq-axis current residual in the dq coordinate system; performing an inverse Park transformation on the residual vector phase angle to obtain an electrical angle value; and mapping the electrical angle value to a specific converter switch to obtain the fault location.

[0048] Specifically, the phase angle of the residual vector of the dq-axis current residual in the dq coordinate system is calculated; an inverse Park transformation is performed on the phase angle of the residual vector to obtain the electrical angle value. It should be understood that although the d-axis and q-axis components of the dq-axis current residual contain fault information, this information is coupled in a synchronously rotating coordinate system, and its values ​​fluctuate periodically with changes in the electrical angle, failing to directly reveal the specific location of the fault in the three-phase physical system. Furthermore, the direction of the spatial vector formed by the dq-axis current residual, i.e., its phase angle, is strongly correlated with the electrical sector where the fault occurred, but this phase angle is relative to the rotating dq coordinate axis, not relative to the fixed three-phase windings. Therefore, in the technical solution of this invention, the phase angle of the residual vector of the dq-axis current residual in the dq coordinate system is further calculated; the phase angle of the residual vector is subjected to an inverse Park transformation to obtain the electrical angle value, thereby first extracting the direction information of the residual vector in the rotating coordinate system. Then, by superimposing it with the phase angle of the grid voltage vector representing the rotation position of the coordinate system itself, the direction information of the extracted residual vector in the rotating coordinate system is restored to the absolute direction information in the stationary coordinate system. In this way, an abstract, dynamically changing fault characteristic can be transformed into a stable electrical angle range directly corresponding to the physical topology of the converter, providing a basis for subsequent precise positioning.

[0049] More specifically, in a specific example of the present invention, firstly, the DSP retrieves the d-axis current residual and q-axis current residual at the current moment, and uses a two-parameter arctangent function to calculate the phase angle of the residual vector in the dq coordinate system, including the d-axis current residual and the q-axis current residual, as the residual vector phase angle; then, the DSP obtains the grid voltage vector phase angle at the same moment from the phase-locked loop module of the wind turbine; next, an addition operation is performed to add the residual vector phase angle to the grid voltage vector phase angle. This operation is conceptually equivalent to the inverse Park transformation, and the result is that the angle of the residual vector is transformed from the rotating coordinate system to the stationary coordinate system; finally, a modulus operation is performed on the added result. The calculations are performed to ensure that the final output electrical angle is normalized to a standard 0 to 1. Within the specified range, the final angle value is the electrical angle value representing the fault location and is passed to the fault mapping process.

[0050] Specifically, the electrical angle value is mapped to a specific converter switch to determine the fault location. It should be understood that the electrical angle value calculated in the previous step is merely a numerical value. While it contains fault location information, it is not a directly usable diagnostic conclusion pointing to a specific physical device and cannot be directly used to guide subsequent control decisions or maintenance operations. Therefore, in the technical solution of this invention, the electrical angle value is further mapped to a specific converter switch to determine the fault location, thereby establishing a deterministic correspondence from the abstract electrical angle to the specific physical switch location. This transforms the complex fault characteristic analysis into a clear and explicit fault location result, such as identifying the specific faulty switch number, thus providing a direct and operable basis for subsequent system fault-tolerant control, alarm information generation, and maintenance guidance.

[0051] More specifically, in a concrete example of the present invention, firstly, during the initialization phase of the converter control system, a mapping table containing six entries is created and stored in the non-volatile memory of the DSP. This mapping table explicitly defines and stores the unique correspondence between six consecutive 60-degree sectors in a 360-degree electrical cycle and six switching transistors (e.g., T1 to T6). Secondly, after the fault location program is triggered and calculates the electrical angle value representing the fault location, the DSP uses this electrical angle value as input. Next, the DSP sequentially compares this electrical angle value with the boundaries of each preset sector interval in the mapping table to determine the sector to which the electrical angle value belongs. Finally, once a matching sector is found, the DSP extracts the transistor identifier corresponding to the sector to which the electrical angle value belongs from the mapping table and outputs the specific converter switching transistor corresponding to the transistor identifier as the final fault location result for system recording or further protection measures.

[0052] In summary, an open-circuit fault detection method for a wind turbine converter switch tube according to an embodiment of the present invention is explained. This method determines whether the system is out of control by comparing the actual output current with the internal command current of the control system, thereby accurately distinguishing between internal faults and external disturbances. First, the residual between the actual current and the reference current on the dq axis is calculated. This residual directly reflects the degree to which the actual behavior deviates from the system's intention. To eliminate the influence of wind speed and load changes on the absolute value of the residual, the residual vector is further normalized using the reference current, constructing a normalized fault characteristic quantity that is consistent across all operating conditions. To improve detection accuracy and avoid misjudgments caused by instantaneous disturbances, the normalized fault characteristic quantity is integrated using a sliding window. Only when the integral value continuously exceeds a preset threshold is the occurrence of the fault confirmed. Once the fault is confirmed, the specific faulty switch tube can be quickly and accurately located using the phase angle information of the current residual vector combined with the grid voltage phase angle.

[0053] This invention also provides a system for detecting open-circuit faults in the switching tubes of a wind turbine converter.

[0054] Figure 5 This is a block diagram of a wind turbine converter switch open-circuit fault detection system according to an embodiment of the present invention. Figure 5 As shown, a wind turbine converter switch open-circuit fault detection system 500 according to an embodiment of the present invention includes: a coordinate transformation module 510, used to perform coordinate transformation on the actual three-phase current output by the wind turbine converter to obtain the actual dq-axis current; a current residual calculation module 520, used to calculate the dq-axis current residual between the actual dq-axis current and the dq-axis reference current; a vector normalization module 530, used to perform residual vector normalization and feature construction on the dq-axis current residual based on the dq-axis reference current to obtain a normalized fault feature quantity; a threshold judgment module 540, used to perform fault feature integration and threshold judgment on the normalized fault feature quantity to obtain a fault identifier; the fault identifier is used to indicate whether a fault is detected; and a fault location module 550, used to perform fault location based on the dq-axis current residual and the grid voltage vector phase angle in response to the fault identifier being true to obtain the fault location.

[0055] The specific implementation method of the wind turbine converter switch tube open circuit fault detection system provided in this embodiment of the invention can be found in the wind turbine converter switch tube open circuit fault detection method provided in this embodiment of the invention, and will not be repeated here.

[0056] A wind turbine converter open-circuit fault detection system 500 according to an embodiment of the present invention can be implemented in various wireless terminals, such as servers with a wind turbine converter open-circuit fault detection control algorithm. In one possible implementation, the wind turbine converter open-circuit fault detection system 500 according to an embodiment of the present invention can be integrated into a wireless terminal as a software module and / or a hardware module. For example, the wind turbine converter open-circuit fault detection system 500 can be a software module in the operating system of the wireless terminal, or it can be an application developed for the wireless terminal; of course, the wind turbine converter open-circuit fault detection system 500 can also be one of many hardware modules of the wireless terminal.

[0057] Those skilled in the art will understand that the above embodiments are specific implementations of the present invention, and in practical applications, various changes can be made in form and detail without departing from the spirit and scope of the present invention.

Claims

1. A method for detecting open-circuit faults in the switching transistors of a wind turbine converter, characterized in that, The method for detecting open-circuit faults in the converter switching transistors of wind turbine generators includes: The actual three-phase current output from the wind turbine converter is transformed by coordinates to obtain the actual current along the dq axis. Calculate the dq-axis current residual between the actual dq-axis current and the dq-axis reference current; Based on the dq-axis reference current, the residual vector of the dq-axis current is normalized and its characteristics are constructed to obtain normalized fault characteristic quantities. The normalized fault feature quantity is integrated with the fault feature quantity and a threshold is determined to obtain the fault identifier; the fault identifier is used to indicate whether a fault has been detected. In response to the fault flag being true, fault location is determined based on the dq-axis current residual and the phase angle of the grid voltage vector.

2. The method for detecting open-circuit faults in the converter switching tubes of a wind turbine generator according to claim 1, characterized in that, The actual three-phase current output from the wind turbine converter is transformed using coordinates to obtain the actual dq-axis current, including: Extract the grid voltage vector phase angle from the phase-locked loop module of the wind turbine; Based on the phase angle of the grid voltage vector, the actual three-phase currents are transformed using the following formula to obtain the actual dq-axis currents: ； in, The phase angle of the grid voltage vector. These are the actual currents of phases A, B, and C, respectively. These are the actual currents along the d-axis and q-axis, respectively.

3. The method for detecting open-circuit faults in the converter switching tubes of a wind turbine generator according to claim 1, characterized in that, Based on the dq-axis reference current, residual vector normalization and feature construction are performed on the dq-axis current residuals to obtain normalized fault characteristic quantities, including: Calculate the reference current for the dq axis reference current; Calculate the magnitude of the current residual vector of the dq-axis current residual; The normalized fault characteristic quantity is obtained by dividing the magnitude of the current residual vector by the reference current quantity.

4. The method for detecting open-circuit faults in the converter switching tubes of a wind turbine generator according to claim 3, characterized in that, The reference current quantity for calculating the dq-axis reference current includes: The reference current for the dq axis is calculated using the following formula: ； in, and These are the d-axis reference current and the q-axis reference current, respectively. This is the reference current.

5. The method for detecting open-circuit faults in the converter switching tubes of a wind turbine generator according to claim 1, characterized in that, Fault identifiers are obtained by integrating normalized fault features and applying threshold judgments, including: Set a sliding time window; Within the sliding time window, the normalized fault feature quantity is integrated to obtain the integrated fault feature. The integrated fault characteristics are compared with a fixed fault threshold to obtain the fault identifier.

6. The method for detecting open-circuit faults in the converter switching tubes of a wind turbine generator according to claim 5, characterized in that, Within the sliding time window, the normalized fault features are integrated to obtain the integrated fault features, including: The fault characteristic integral is performed on the normalized fault characteristic quantity using the following formula: ； in, To normalize fault characteristic quantities, For integration variables, For sliding time windows, for Integral fault characteristics at any given time.

7. The method for detecting open-circuit faults in the converter switching tubes of a wind turbine generator according to claim 5, characterized in that, Fault identifiers are obtained by comparing integral fault characteristics with fixed fault thresholds, including: In response to the duration for which the integral fault characteristic exceeds a fixed fault threshold for a predetermined period of time, the fault flag is set to true.

8. The method for detecting open-circuit faults in the converter switching tubes of a wind turbine generator according to claim 1, characterized in that, In response to the fault flag being true, fault location is determined based on the dq-axis current residual and the grid voltage vector phase angle, including: Calculate the phase angle of the residual vector of the dq-axis current residual in the dq coordinate system; The phase angle of the residual vector is subjected to inverse Park transformation to obtain the electrical angle value; The electrical angle value is mapped to a specific converter switch to determine the fault location.

9. The method for detecting open-circuit faults in the converter switching tubes of a wind turbine generator according to claim 8, characterized in that, Mapping the electrical angle range to specific converter switching transistors to obtain the fault location includes: According to the preset mapping table, the specific converter switch corresponding to the electrical angle value is taken as the fault location; The preset mapping table stores the unique correspondence between six consecutive 60-degree sectors in a 360-degree electrical cycle and six switching transistors.

10. A system for detecting open-circuit faults in the switching transistors of a wind turbine converter, characterized in that, The wind turbine converter switch tube open circuit fault detection system includes: The coordinate transformation module is used to perform coordinate transformation on the actual three-phase current output by the wind turbine converter to obtain the actual current of the dq axis. The current residual calculation module is used to calculate the dq-axis current residual between the actual dq-axis current and the dq-axis reference current. The vector normalization module is used to perform residual vector normalization and feature construction on the dq axis current residual based on the dq axis reference current, so as to obtain normalized fault feature quantities. The threshold determination module is used to perform fault feature integration and threshold determination on the normalized fault feature quantity to obtain a fault identifier; the fault identifier is used to indicate whether a fault has been detected. The fault location module is used to locate the fault position based on the dq-axis current residual and the phase angle of the grid voltage vector in response to the fault identifier being true.

Images