Online monitoring system and method for wind turbine generator blades, electronic equipment and medium

Abstract

The invention provides an online monitoring system and method for a wind turbine generator blade, electronic equipment and a medium, and belongs to the field of wind turbine generator testing. The system comprises a low-frequency vibration sensor which is installed at the position, away from a blade root by a preset distance, of a wind turbine generator blade and used for collecting vibration signals of the wind turbine generator blade; the CANopen communication module is used for receiving and transmitting a vibration signal of the wind turbine generator blade; the CANopen master station terminal module is used for transmitting the vibration signals of the blades of the wind turbine generator to the master control system; and the main control system processes and analyzes the blade vibration signals through an embedded fault diagnosis algorithm, and realizes real-time monitoring of the blades of the wind turbine generator. By establishing a communication link between the blade sensor and the master control system, low-frequency vibration signals of the blade can be collected, a fault diagnosis algorithm is embedded, real-time blade impact monitoring and safety protection are achieved, and the blade impact monitoring system has the advantages of being high in diagnosis coverage rate and long in average failure occurrence time.

Description

Technical Field

[0001] The invention belongs to the technical field of wind turbine generator set testing, and in particular relates to an online monitoring system, method, electronic equipment and medium for wind turbine generator set blades. Background Art

[0002] In recent years, the development of super-large wind turbines has become a trend in the wind power industry. In order to improve the utilization rate of wind energy and increase power generation, the mass and length of wind turbine blades have continued to increase, and the impact of blades on the safety and stability of wind turbine operation is particularly obvious. In addition, due to surface or structural damage, icing, lightning damage, or sudden changes in wind speed and wind direction, the blades of wind turbines vibrate too much, the load exceeds the design limit, and the phenomenon of breaking or sweeping the tower also occurs from time to time. Therefore, it is particularly important for the online monitoring system of wind turbine blades to monitor the health of the blades, timely detect and warn potential faults, and improve the operation and maintenance efficiency and safety of the unit.

[0003] At present, the online monitoring system of wind turbine blades is an auxiliary control system and does not participate in the control of the unit. When the unit has sudden faults such as blade breakage or tower sweeping, it is impossible to control the unit to quickly perform actions such as shutdown. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide an online monitoring system, method, electronic equipment and medium for wind turbine blades, which are used to completely or at least partially solve the technical problems existing in the above-mentioned prior art that potential blade failures cannot be detected in time and the unit cannot be controlled to shut down when a failure occurs.

[0005] In a first aspect, an embodiment of the present application provides an online monitoring system for a blade of a wind turbine generator set, comprising: A low-frequency vibration sensor is installed at a preset distance between the blade of the wind turbine and the blade root to collect vibration signals of the blade of the wind turbine; A CANopen communication module, connected to the low-frequency vibration sensor, for receiving and transmitting vibration signals of the wind turbine blades; A CANopen master terminal module, which is communicatively connected to the CANopen communication module and is used to transmit the vibration signal of the wind turbine blade to the main control system; The main control system processes and analyzes blade vibration signals through an embedded fault diagnosis algorithm to achieve real-time monitoring of wind turbine blades.

[0006] Optionally, a 24VDC power supply lightning protection module is added to the CANopen communication module to protect the power supply from being struck by lightning; A CANopen lightning protection module is added to the CANopen master terminal module to protect the main control system from being struck by lightning.

[0007] Optional, CANopen lightning protection module is used for 6V high-speed signal line, with voltage protection level less than 25V and lightning impulse current 5kA, for CAN network lightning protection; The DC working voltage of the 24VDC power supply lightning protection module is 24VDC, and the voltage protection level is less than or equal to 250V. It is used for 24VDC power supply lightning protection.

[0008] Optionally, the low-frequency vibration sensor is a capacitive dual-axis accelerometer with a narrow-band filter.

[0009] Optionally, the main control system processes and analyzes blade vibration signals through an embedded fault diagnosis algorithm to achieve real-time monitoring of wind turbine blades, including: By monitoring the blade surface damage index, it is judged whether the blade has surface damage and its severity; by the blade structure damage index, it is judged whether the blade has structural damage and its severity; by the blade imbalance index, it is judged whether the blade is unbalanced and its severity; by the blade icing monitoring index, it is judged whether the blade has ice; by the natural frequency monitoring index, the first-order and second-order natural frequencies of the blade flapping and swinging directions are characterized; by the data similarity index, the similarity of the operating conditions of the three blades is characterized.

[0010] Optionally, the process of determining the blade surface damage index includes two calculation methods: spectrum centroid and differential dispersion: Use librosa.feature.spectral_centroid to calculate the spectral centroid of each frame of the vibration signal and find the median to obtain the first index value; After using the time domain waveform file for differential discrete transformation, the standard deviation is used to measure the discrete degree to obtain the second index value; Combining the spectrum centroid and differential dispersion of the blade surface damage index, the monthly median is compared with the preset threshold to determine the ratio of the first index value and the second index value exceeding the preset threshold, and the maximum value of the three leaves is taken as the blade surface damage index; The process of determining the blade structure damage index includes: measuring from the theoretical change rate and relative change rate of the swing stiffness: The theoretical change rate of the shimmy stiffness is determined by the ratio of the peak position of the blade shimmy natural first-order frequency obtained through the amplitude spectrum diagram and the actual shimmy natural first-order frequency of the blade; The relative change rate of the shimmy stiffness is determined by the ratio of the peak position of the first-order frequency of the shimmy inherent of a blade obtained through the amplitude spectrum diagram to the average of the peak positions obtained by the remaining two blades; Combining the theoretical change rate and relative change rate of the shimmy stiffness, the monthly median is compared with the preset threshold to determine the proportion of the theoretical change rate and relative change rate of the shimmy stiffness exceeding the threshold, and the maximum value of the three blades is taken as the blade structure damage index; The process of determining the blade imbalance index includes: The flapping signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the flapping imbalance index; The shimmy signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the shimmy imbalance index; Combine the flapping imbalance index and the shimmy imbalance index, compare the monthly median with the preset threshold, and take the proportion of the flapping imbalance index and the shimmy imbalance index exceeding the preset threshold as the blade imbalance index; The process of determining blade icing monitoring indicators includes: Extract the blade temperature value in the working condition file, where the external environmental condition for blade icing is a temperature value below degrees; The process of determining the blade natural frequency monitoring index includes: the effective value of the first-order shimmy and the peak position of the first-order shimmy: Perform RMS statistics on the spectrum amplitude near the first-order natural frequency of the shimmy vibration, extract the peak position near the first-order natural frequency of the shimmy vibration, compare the monthly median with the threshold, and use the ratio of the effective value of the first-order shimmy vibration and the peak position of the first-order shimmy vibration exceeding the preset threshold as the blade natural frequency monitoring index; The process of determining the leaf data similarity index includes: The cosine similarity between the three blades of the amplitude spectrum is measured, and then the minimum similarity is statistically calculated to determine the frequency domain similarity; Calculate the time domain index, measure the cosine similarity between the three leaves of the time domain index sequence, and then calculate the minimum similarity to determine the time domain similarity; The data of the time domain waveform file is first processed by first-order difference, and then the time domain index value is counted, and then the cosine similarity between the three leaves of the sequence of the time domain index is measured, and then the minimum similarity is counted to determine the time domain difference similarity; The three types of indicators, namely frequency domain, time domain and time domain difference, are combined. The monthly median is compared with the preset threshold, and the proportion of the three types of indicators, namely frequency domain, time domain and time domain difference, exceeding the preset threshold is used as the leaf data similarity index. In a second aspect, an embodiment of the present application further provides an online monitoring method for wind turbine blades applied to the online monitoring system for wind turbine blades, comprising: The vibration signals of the wind turbine blades are collected through low-frequency vibration sensors; The blade vibration signal is processed and analyzed using a preset fault diagnosis algorithm to achieve real-time monitoring of wind turbine blades.

[0011] Optionally, a preset fault diagnosis algorithm is used to process and analyze blade vibration signals to achieve real-time monitoring of wind turbine blades, including: Through the monitored blade surface damage index, it is judged whether the blade has surface damage and its severity; through the blade structure damage index, it is judged whether the blade has structural damage and its severity; through the blade imbalance index, it is judged whether the blade is unbalanced and its severity; through the blade icing monitoring index, it is judged whether the blade has ice; through the natural frequency monitoring index, the first-order and second-order natural frequencies of the blade in the flapping and swinging directions are characterized; through the data similarity index, the similarity of the operating conditions of the three blades is characterized; among them, The determination process of blade surface damage index includes two calculation methods: spectrum centroid and differential dispersion: Use librosa.feature.spectral_centroid to calculate the spectral centroid of each frame of the vibration signal and find the median to obtain the first index value; After using the time domain waveform file for differential discrete transformation, the standard deviation is used to measure the discrete degree to obtain the second index value; Combining the spectrum centroid and differential dispersion of the blade surface damage index, the monthly median is compared with the preset threshold to determine the ratio of the first index value and the second index value exceeding the preset threshold, and the maximum value of the three leaves is taken as the blade surface damage index; The process of determining the blade structure damage index includes: measuring from the theoretical change rate and relative change rate of the swing stiffness: The theoretical change rate of the shimmy stiffness is determined by the ratio of the peak position of the blade shimmy natural first-order frequency obtained through the amplitude spectrum diagram and the actual shimmy natural first-order frequency of the blade; The relative change rate of the shimmy stiffness is determined by the ratio of the peak position of the first-order frequency of the shimmy inherent of a blade obtained through the amplitude spectrum diagram to the average of the peak positions obtained by the remaining two blades; Combining the theoretical change rate and relative change rate of the shimmy stiffness, the monthly median is compared with the preset threshold to determine the proportion of the theoretical change rate and relative change rate of the shimmy stiffness exceeding the threshold, and the maximum value of the three blades is taken as the blade structure damage index; The process of determining the blade imbalance index includes: The flapping signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the flapping imbalance index; The shimmy signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the shimmy imbalance index; Combine the flapping imbalance index and the shimmy imbalance index, compare the monthly median with the preset threshold, and take the proportion of the flapping imbalance index and the shimmy imbalance index exceeding the preset threshold as the blade imbalance index; The process of determining blade icing monitoring indicators includes: Extract the blade temperature value in the working condition file, where the external environmental condition for blade icing is a temperature value below degrees; The process of determining the blade natural frequency monitoring index includes: the effective value of the first-order shimmy and the peak position of the first-order shimmy: Perform RMS statistics on the spectrum amplitude near the first-order natural frequency of the shimmy vibration, extract the peak position near the first-order natural frequency of the shimmy vibration, compare the monthly median with the threshold, and use the ratio of the effective value of the first-order shimmy vibration and the peak position of the first-order shimmy vibration exceeding the preset threshold as the blade natural frequency monitoring index; The process of determining the leaf data similarity index includes: The cosine similarity between the three blades of the amplitude spectrum is measured, and then the minimum similarity is statistically calculated to determine the frequency domain similarity; Calculate the time domain index, measure the cosine similarity between the three leaves of the time domain index sequence, and then calculate the minimum similarity to determine the time domain similarity; The data of the time domain waveform file is first processed by first-order difference, and then the time domain index value is counted, and then the cosine similarity between the three leaves of the sequence of the time domain index is measured, and then the minimum similarity is counted to determine the time domain difference similarity; The three types of indicators, namely frequency domain, time domain and time domain difference, are combined. The monthly median is compared with the preset threshold, and the proportion of the three types of indicators, namely frequency domain, time domain and time domain difference, exceeding the preset threshold is used as the leaf data similarity index. In a third aspect, an embodiment of the present application further provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned online monitoring method for wind turbine blades when executing the program.

[0012] In a fourth aspect, an embodiment of the present application further provides a storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of the above-mentioned method for online monitoring of wind turbine blades.

[0013] It can be seen from the above technical solutions that the present invention has the following advantages: In the online monitoring system, method, electronic equipment and medium for wind turbine blades provided in the present application, by establishing a communication link between the blade sensor and the PLC main control system, the low-frequency vibration signal of the blade can be collected, and a fault diagnosis algorithm can be embedded to achieve real-time monitoring and safety protection of blade impact, with high diagnostic coverage and a long average failure time. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solution of the present invention, the accompanying drawings required for use in the description will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0015] Figure 1 A schematic structural diagram of an online monitoring system for wind turbine blades provided by an embodiment of the present invention; Figure 2 An implementation flow chart of an online monitoring method for wind turbine blades provided by an embodiment of the present invention applied to an online monitoring system for wind turbine blades; Figure 3 The figure is a schematic diagram of the hardware structure of an electronic device according to various embodiments of the present invention. DETAILED DESCRIPTION

[0016] In the detailed description below, various embodiments of the present disclosure will be described more fully. The present disclosure may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of the present disclosure to the specific embodiments disclosed herein, but rather the present disclosure should be understood to cover all adjustments, equivalents and / or alternatives that fall within the spirit and scope of the various embodiments of the present disclosure.

[0017] Hereinafter, the terms "include" or "may include" that may be used in various embodiments of the present disclosure indicate the presence of the disclosed functions or operations, and do not limit the addition of one or more functions or operations. In addition, as used in various embodiments of the present disclosure, the terms "include", "have" and their cognates are intended only to indicate specific features, numbers, steps, operations, or combinations of the foregoing items, and should not be understood as first excluding the presence of one or more other features, numbers, steps, operations, or combinations of the foregoing items or the possibility of adding one or more features, numbers, steps, operations, or combinations of the foregoing items.

[0018] In various embodiments of the present disclosure, the expression "or" or "at least one of A or / and B" includes any combination or all combinations of the words listed at the same time. For example, the expression "A or B" or "at least one of A or / and B" may include A, may include B, or may include both A and B.

[0019] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0020] See also Figure 1 The figure is a schematic diagram of the structure of an online monitoring system for wind turbine blades in a specific embodiment, including: A low-frequency vibration sensor is installed at a preset distance between the blade of the wind turbine and the blade root to collect vibration signals of the blade of the wind turbine; A CANopen communication module, connected to the low-frequency vibration sensor, for receiving and transmitting vibration signals of the wind turbine blades; A CANopen master terminal module, which is communicatively connected to the CANopen communication module and is used to transmit the vibration signal of the wind turbine blade to the main control system; The main control system processes and analyzes blade vibration signals through an embedded fault diagnosis algorithm to achieve real-time monitoring of wind turbine blades.

[0021] Exemplarily, a low-frequency vibration sensor is installed on the wind turbine blade at a position 1 / 3 away from the blade root to collect vibration signals in two directions: blade flapping and swinging (flashing signal refers to the signal of bending vibration in the direction perpendicular to the rotation plane; swinging signal refers to the signal of bending vibration in the rotation plane), and the main control system is a PLC main control system.

[0022] In a specific implementation, a 24VDC power supply lightning protection module is added to the CANopen communication module to protect the power supply from being struck by lightning; a CANopen lightning protection module is added to the CANopen master terminal module to protect the main control system from being struck by lightning, wherein the CANopen lightning protection module is applied to a 6V high-speed signal line, with a voltage protection level of less than 25V and a lightning impulse current of 5kA, and is used for lightning protection of the CAN network; the 24VDC power supply lightning protection module has a DC working voltage of 24VDC and a voltage protection level of less than or equal to 250V, and is used for lightning protection of the 24VDC power supply.

[0023] In this embodiment, the type of low-frequency vibration sensor is a capacitive dual-axis accelerometer with a narrowband filter, a frequency range of 0.05-1000Hz, and detectors: safety shock detection (SSD), effective value, peak value, peak-to-peak value. Through redundant safety relays, system faults can be output to the control system.

[0024] Preferably, the CANopen communication module has a 4-channel electrically isolated CAN interface, and the communication baud rate can be configured through the RS232 port: 5Kbps-1Mbps, which is used to increase the number of nodes in the CAN network and change the topology of the CAN network wiring; the CANopen master terminal module is a CANopen gateway for CANopen protocol conversion.

[0025] In a specific embodiment, Figure 1 It includes a pitch cabinet, a terminal block (2 in and 8 out), a low-frequency vibration sensor (including low-frequency sensors A, B, and C with 120Ω terminal resistance), a CANopen communication module (including a 24VDC power supply lightning protection module), a slip ring communication cable, a CANopen master terminal module (including a CANopen lightning protection module), and a PLC main control system. The red arrow indicates the power supply 24V+, the blue arrow indicates the power supply 24V-, and the green arrow indicates CAN high level, CAN low level, and CAN grounding.

[0026] Its working principle is as follows: three blade low-frequency vibration sensors are installed on the blade at a position 1 / 3 away from the blade root. They can collect vibration signals in two directions: blade flapping and swinging. The low-frequency vibration signal is transmitted to the CANopen communication module. In order to avoid lightning strikes, a 24VDC power supply lightning protection module is added to protect the power supply. The signal is transmitted to the PLC main control system through the slip ring communication cable and the CANopen master station terminal module. In order to avoid lightning strikes, a CANopen lightning protection module is added to protect the PLC main control system. The PLC main control system is embedded with fault diagnosis algorithms such as blade surface or structural damage, blade imbalance, blade icing, natural frequency, and data similarity to realize real-time monitoring of blade impact.

[0027] In a specific implementation, the main control system processes and analyzes blade vibration signals through an embedded fault diagnosis algorithm to achieve real-time monitoring of wind turbine blades, including: By monitoring the blade surface damage index, it is judged whether the blade has surface damage and its severity; by the blade structure damage index, it is judged whether the blade has structural damage and its severity; by the blade imbalance index, it is judged whether the blade is unbalanced and its severity; by the blade icing monitoring index, it is judged whether the blade has ice; by the natural frequency monitoring index, the first-order and second-order natural frequencies of the blade flapping and swinging directions are characterized; by the data similarity index, the similarity of the operating conditions of the three blades is characterized.

[0028] Specifically, the process of determining the blade surface damage index includes two calculation methods: spectrum centroid and differential dispersion: Use librosa.feature.spectral_centroid to calculate the spectral centroid of each frame of the vibration signal, and find the median to obtain the first index value; After using the time domain waveform file for differential discrete transformation, the standard deviation is used to measure the discrete degree to obtain the second index value; Combining the spectrum centroid and differential dispersion of the blade surface damage index, the monthly median is compared with the preset threshold to determine the ratio of the first index value and the second index value exceeding the preset threshold, and the maximum value of the three leaves is taken as the blade surface damage index; The process of determining the blade structure damage index includes: measuring from the theoretical change rate and relative change rate of the swing stiffness: The theoretical change rate of the shimmy stiffness is determined by the ratio of the peak position of the blade shimmy natural first-order frequency obtained through the amplitude spectrum diagram and the actual shimmy natural first-order frequency of the blade; The relative change rate of the shimmy stiffness is determined by the ratio of the peak position of the first-order frequency of the shimmy inherent of a blade obtained through the amplitude spectrum diagram to the average of the peak positions obtained by the remaining two blades; Combining the theoretical change rate and relative change rate of the shimmy stiffness, the monthly median is compared with the preset threshold to determine the proportion of the theoretical change rate and relative change rate of the shimmy stiffness exceeding the threshold, and the maximum value of the three blades is taken as the blade structure damage index; The process of determining the blade imbalance index includes: The flapping signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the flapping imbalance index; The shimmy signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the shimmy imbalance index; Combine the flapping imbalance index and the shimmy imbalance index, compare the monthly median with the preset threshold, and take the proportion of the flapping imbalance index and the shimmy imbalance index exceeding the preset threshold as the blade imbalance index; The process of determining blade icing monitoring indicators includes: Extract the blade temperature value in the working condition file, where the external environmental condition for blade icing is a temperature value below degrees; The process of determining the blade natural frequency monitoring index includes: the effective value of the first-order shimmy and the peak position of the first-order shimmy: Perform RMS statistics on the spectrum amplitude near the first-order natural frequency of the shimmy vibration, extract the peak position near the first-order natural frequency of the shimmy vibration, compare the monthly median with the threshold, and use the ratio of the effective value of the first-order shimmy vibration and the peak position of the first-order shimmy vibration exceeding the preset threshold as the blade natural frequency monitoring index; The process of determining the leaf data similarity index includes: The cosine similarity between the three blades of the amplitude spectrum is measured, and then the minimum similarity is statistically calculated to determine the frequency domain similarity; Calculate the time domain index, measure the cosine similarity between the three leaves of the time domain index sequence, and then calculate the minimum similarity to determine the time domain similarity; The data of the time domain waveform file is first processed by first-order difference, and then the time domain index value is counted, and then the cosine similarity between the three leaves of the sequence of the time domain index is measured, and then the minimum similarity is counted to determine the time domain difference similarity; The three types of indicators, namely frequency domain, time domain and time domain difference, are combined. The monthly median is compared with the preset threshold, and the proportion of the three types of indicators, namely frequency domain, time domain and time domain difference, exceeding the preset threshold is used as the leaf data similarity index. like Figure 2 As shown, the following is an embodiment of the online monitoring method for wind turbine blades provided by the embodiment of the present disclosure and applied to the online monitoring system for wind turbine blades, which belongs to the same inventive concept as the online monitoring system for wind turbine blades in the above-mentioned embodiments. For details not described in detail in the embodiment of the online monitoring method for wind turbine blades, please refer to the embodiment of the online monitoring system for wind turbine blades mentioned above.

[0029] The PLC main control system is embedded with a blade fault diagnosis algorithm, and the vibration data in step C is processed and analyzed. The characteristic indicators monitored by the system include blade surface damage indicators, which are used to judge whether the blade has surface damage and its severity; blade structural damage indicators, which are used to judge whether the blade has structural damage and its severity; blade imbalance indicators, which are used to judge whether the blade has imbalance and its severity; blade icing monitoring indicators, which are used to judge whether the blade has icing; natural frequency monitoring indicators, which are used to characterize the first-order and second-order natural frequencies of the blade in the flapping and swinging directions; and data similarity indicators, which are used to characterize the similarity of the operating states of the three blades. The PLC main control system can realize batch offline data import, display fault indicators and intelligent diagnosis results, and the output data should at least contain characteristic data. The output display graphics include trend charts, time domain charts, and spectrum charts. The pre-processing functions include time filtering, data deletion, and data checking. Operations include sideband cursors, multiple frequency cursors, zooming in, zooming out, and clicking data to display amplitude and time for various types of spectra.

[0030] The specific steps include the following: Step 200: collecting vibration signals of wind turbine blades through a low-frequency vibration sensor.

[0031] For example, taking a wind farm in Liaoning Province as an example, the wind farm has 24 3.XMW units put into operation, so one unit is selected for online real-time monitoring of blades and hanging test of the safety protection system. The low-frequency vibration sensor is installed on the blade of the wind turbine at a position 1 / 3 away from the blade root, and can collect vibration signals in two directions of blade flapping and swinging. The sensor type is a capacitive dual-axis accelerometer with a narrowband filter, a frequency range of 0.05-1000Hz, and detectors: safety shock detection (SSD), effective value, peak value, peak-peak value. Through redundant safety relays, system faults are output to the control system; the CANopen communication module has a 4-channel electrically isolated CAN interface, and the communication baud rate can be configured through the RS232 port: 5Kbps-1Mbps; the CANopen master terminal module is a CANopen gateway for CANopen protocol conversion; the PLC master control system can connect and control CANopen terminal equipment (low-frequency vibration sensor) through the CANopen master terminal module; the CANopen lightning protection module is applied to 6V high-speed signal lines, with a voltage protection level of <25V and a lightning impulse current of 5kA; the DC working voltage of the 24VDC power supply lightning protection module is 24VDC, and the protection level is ≤250V.

[0032] First, the online real-time monitoring and safety protection system for wind turbine blades is installed, including installing low-frequency vibration sensors at 1 / 3 of the distance from the root of the three blades, installing the sensor cables in an "S" shape, routing the cables at the position of the blade cover, installing the mechanical interface, installing the electrical interface, etc.; then the sensor is powered by 24VDC of the pitch control cabinet, with a built-in 120Ω terminal resistor, collecting vibration signals in the two directions of blade swing and shimmy, and then using the CANopen communication protocol, the signal is transmitted to the CANopen communication module, through the slip ring communication cable, and transmitted to the PLC main control system through the CANopen master terminal module. The vibration data collected by the low-frequency vibration sensor mainly includes the original data of blades 1, 2, and 3, the frequency band data of blades 1, 2, and 3, and the status data of blades 1, 2, and 3.

[0033] Step 201: Utilize a preset fault diagnosis algorithm to process and analyze blade vibration signals to achieve real-time monitoring of wind turbine blades.

[0034] Exemplarily, the PLC main control system is embedded with a blade fault diagnosis algorithm to process and analyze vibration data. The characteristic indicators monitored by the system include blade surface damage indicators, which are used to judge whether the blade has surface damage and its severity; blade structural damage indicators, which are used to judge whether the blade has structural damage and its severity; blade imbalance indicators, which are used to judge whether the blade has imbalance and its severity; blade icing monitoring indicators, which are used to judge whether the blade has ice; natural frequency monitoring indicators, which are used to characterize the first-order and second-order natural frequencies in the blade flapping and swinging directions; and data similarity indicators, which are used to characterize the similarity of the operating states of the three blades. The PLC main control system can realize batch offline data import, display fault indicators and intelligent diagnosis results, and the output data should at least contain characteristic data. The output display graphics include trend charts, time domain charts, and spectrum charts. The pre-processing functions include time filtering, data deletion, and data selection. Operations include sideband cursors, multiple frequency cursors, zooming in, zooming out, and clicking data to display amplitude and time for various types of spectra.

[0035] Specifically, when executing step 201, the following steps may be specifically performed: Through the monitored blade surface damage index, it is judged whether the blade has surface damage and its severity; through the blade structure damage index, it is judged whether the blade has structural damage and its severity; through the blade imbalance index, it is judged whether the blade is unbalanced and its severity; through the blade icing monitoring index, it is judged whether the blade has ice; through the natural frequency monitoring index, the first-order and second-order natural frequencies of the blade in the flapping and swinging directions are characterized; through the data similarity index, the similarity of the operating conditions of the three blades is characterized; among them, The determination process of blade surface damage index includes two calculation methods: spectrum centroid and differential dispersion: Use librosa.feature.spectral_centroid to calculate the spectral centroid of each frame of the vibration signal and find the median to obtain the first index value; After using the time domain waveform file for differential discrete transformation, the standard deviation is used to measure the discrete degree to obtain the second index value; Combining the spectrum centroid and differential dispersion of the blade surface damage index, the monthly median is compared with the preset threshold to determine the ratio of the first index value and the second index value exceeding the preset threshold, and the maximum value of the three leaves is taken as the blade surface damage index; The process of determining the blade structure damage index includes: measuring from the theoretical change rate and relative change rate of the swing stiffness: The theoretical change rate of the shimmy stiffness is determined by the ratio of the peak position of the blade shimmy natural first-order frequency obtained through the amplitude spectrum diagram and the actual shimmy natural first-order frequency of the blade; The relative change rate of the shimmy stiffness is determined by the ratio of the peak position of the first-order frequency of the shimmy inherent of a blade obtained through the amplitude spectrum diagram to the average of the peak positions obtained by the remaining two blades; Combining the theoretical change rate and relative change rate of the shimmy stiffness, the monthly median is compared with the preset threshold to determine the proportion of the theoretical change rate and relative change rate of the shimmy stiffness exceeding the threshold, and the maximum value of the three blades is taken as the blade structure damage index; The process of determining the blade imbalance index includes: The flapping signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the flapping imbalance index; The shimmy signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the shimmy imbalance index; Combine the flapping imbalance index and the shimmy imbalance index, compare the monthly median with the preset threshold, and take the proportion of the flapping imbalance index and the shimmy imbalance index exceeding the preset threshold as the blade imbalance index; The process of determining blade icing monitoring indicators includes: Extract the blade temperature value in the working condition file, where the external environmental condition for blade icing is a temperature value below degrees; The process of determining the blade natural frequency monitoring index includes: the effective value of the first-order shimmy and the peak position of the first-order shimmy: Perform RMS statistics on the spectrum amplitude near the first-order natural frequency of the shimmy vibration, extract the peak position near the first-order natural frequency of the shimmy vibration, compare the monthly median with the threshold, and use the ratio of the effective value of the first-order shimmy vibration and the peak position of the first-order shimmy vibration exceeding the preset threshold as the blade natural frequency monitoring index; The process of determining the leaf data similarity index includes: The cosine similarity between the three blades of the amplitude spectrum is measured, and then the minimum similarity is statistically calculated to determine the frequency domain similarity; Calculate the time domain index, measure the cosine similarity between the three leaves of the time domain index sequence, and then calculate the minimum similarity to determine the time domain similarity; The data of the time domain waveform file is first processed by first-order difference, and then the time domain index value is counted, and then the cosine similarity between the three leaves of the sequence of the time domain index is measured, and then the minimum similarity is counted to determine the time domain difference similarity; The three types of indicators, namely frequency domain, time domain and time domain difference, are combined. The monthly median is compared with the preset threshold, and the proportion of the three types of indicators, namely frequency domain, time domain and time domain difference, exceeding the preset threshold is used as the leaf data similarity index.

[0036] By establishing a communication link between the blade sensor and the PLC main control system, the blade's low-frequency vibration signal can be collected, and fault diagnosis algorithms such as blade surface or structural damage, blade imbalance, blade icing, natural frequency, and data similarity can be embedded. The purpose is to achieve real-time monitoring and safety protection of blade impact, with high diagnostic coverage and a long average failure time.

[0037] Figure 3 It is a schematic diagram of the hardware structure of an electronic device for implementing various embodiments of the present invention.

Claims

1. An online monitoring system for wind turbine blades, characterized in that: include: A low-frequency vibration sensor is installed at a preset distance between the blade of the wind turbine and the blade root to collect vibration signals of the blade of the wind turbine; A CANopen communication module, connected to the low-frequency vibration sensor, for receiving and transmitting vibration signals of the wind turbine blades; A CANopen master terminal module, which is communicatively connected to the CANopen communication module and is used to transmit the vibration signal of the wind turbine blade to the main control system; The main control system processes and analyzes blade vibration signals through an embedded fault diagnosis algorithm to achieve real-time monitoring of wind turbine blades.

2. The online monitoring system for wind turbine blades according to claim 1, characterized in that: A 24VDC power supply lightning protection module is added to the CANopen communication module to protect the power supply from being struck by lightning; A CANopen lightning protection module is added to the CANopen master terminal module to protect the main control system from being struck by lightning.

3. The online monitoring system for wind turbine blades according to claim 2, characterized in that: CANopen lightning protection module is used for 6V high-speed signal line, with voltage protection level less than 25V and lightning impulse current of 5kA, and is used for lightning protection of CAN network; The DC working voltage of the 24VDC power supply lightning protection module is 24VDC, and the voltage protection level is less than or equal to 250V. It is used for 24VDC power supply lightning protection.

4. The online monitoring system for wind turbine blades according to claim 1, characterized in that: The low-frequency vibration sensor is a capacitive dual-axis accelerometer with a narrow-band filter.

5. The online monitoring system for wind turbine blades according to claim 4, characterized in that: The main control system processes and analyzes blade vibration signals through an embedded fault diagnosis algorithm to achieve real-time monitoring of wind turbine blades, including: By monitoring the blade surface damage index, it is judged whether the blade has surface damage and its severity; by the blade structure damage index, it is judged whether the blade has structural damage and its severity; by the blade imbalance index, it is judged whether the blade is unbalanced and its severity; by the blade icing monitoring index, it is judged whether the blade has ice; by the natural frequency monitoring index, the first-order and second-order natural frequencies of the blade flapping and swinging directions are characterized; by the data similarity index, the similarity of the operating conditions of the three blades is characterized.

6. The online monitoring system for wind turbine blades according to claim 5, characterized in that: The determination process of blade surface damage index includes two calculation methods: spectrum centroid and differential dispersion: Use librosa.feature.spectral_centroid to calculate the spectral centroid of each frame of the vibration signal and find the median to obtain the first index value; After using the time domain waveform file for differential discrete transformation, the standard deviation is used to measure the discrete degree to obtain the second index value; Combining the spectrum centroid and differential dispersion of the blade surface damage index, the monthly median is compared with the preset threshold to determine the ratio of the first index value and the second index value exceeding the preset threshold, and the maximum value of the three leaves is taken as the blade surface damage index; The process of determining the blade structure damage index includes: measuring from the theoretical change rate and relative change rate of the swing stiffness: The theoretical change rate of the shimmy stiffness is determined by the ratio of the peak position of the blade shimmy natural first-order frequency obtained through the amplitude spectrum diagram and the actual shimmy natural first-order frequency of the blade; The relative change rate of the shimmy stiffness is determined by the ratio of the peak position of the first-order frequency of the shimmy inherent of a blade obtained through the amplitude spectrum diagram to the average of the peak positions obtained by the remaining two blades; Combining the theoretical change rate and relative change rate of the shimmy stiffness, the monthly median is compared with the preset threshold to determine the proportion of the theoretical change rate and relative change rate of the shimmy stiffness exceeding the threshold, and the maximum value of the three blades is taken as the blade structure damage index; The process of determining the blade imbalance index includes: The flapping signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the flapping imbalance index; The shimmy signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the shimmy imbalance index; Combine the flapping imbalance index and the shimmy imbalance index, compare the monthly median with the preset threshold, and take the proportion of the flapping imbalance index and the shimmy imbalance index exceeding the preset threshold as the blade imbalance index; The process of determining blade icing monitoring indicators includes: Extract the blade temperature value in the working condition file, where the external environmental condition for blade icing is a temperature value below degrees; The process of determining the blade natural frequency monitoring index includes: the effective value of the first-order shimmy and the peak position of the first-order shimmy: Perform RMS statistics on the spectrum amplitude near the first-order natural frequency of the shimmy vibration, extract the peak position near the first-order natural frequency of the shimmy vibration, compare the monthly median with the threshold, and use the ratio of the effective value of the first-order shimmy vibration and the peak position of the first-order shimmy vibration exceeding the preset threshold as the blade natural frequency monitoring index; The process of determining the leaf data similarity index includes: The cosine similarity between the three blades of the amplitude spectrum is measured, and then the minimum similarity is statistically calculated to determine the frequency domain similarity; Calculate the time domain index, measure the cosine similarity between the three leaves of the time domain index sequence, and then calculate the minimum similarity to determine the time domain similarity; The data of the time domain waveform file is first processed by first-order difference, and then the time domain index value is counted, and then the cosine similarity between the three leaves of the sequence of the time domain index is measured, and then the minimum similarity is counted to determine the time domain difference similarity; The three types of indicators, namely frequency domain, time domain and time domain difference, are combined. The monthly median is compared with the preset threshold, and the proportion of the three types of indicators, namely frequency domain, time domain and time domain difference, exceeding the preset threshold is used as the leaf data similarity index.

7. An online monitoring method for wind turbine blades applied to the online monitoring system for wind turbine blades according to any one of claims 1 to 6, characterized in that: include: The vibration signals of the wind turbine blades are collected through low-frequency vibration sensors; The blade vibration signal is processed and analyzed using a preset fault diagnosis algorithm to achieve real-time monitoring of wind turbine blades.

8. The online monitoring method for wind turbine blades according to claim 7, characterized in that: The blade vibration signal is processed and analyzed using the preset fault diagnosis algorithm to achieve real-time monitoring of wind turbine blades, including: Through the monitored blade surface damage index, it is judged whether the blade has surface damage and its severity; through the blade structure damage index, it is judged whether the blade has structural damage and its severity; through the blade imbalance index, it is judged whether the blade is unbalanced and its severity; through the blade icing monitoring index, it is judged whether the blade has ice; through the natural frequency monitoring index, the first-order and second-order natural frequencies of the blade in the flapping and swinging directions are characterized; through the data similarity index, the similarity of the operating conditions of the three blades is characterized; among them, The determination process of blade surface damage index includes two calculation methods: spectrum centroid and differential dispersion: Use librosa.feature.spectral_centroid to calculate the spectral centroid of each frame of the vibration signal and find the median to obtain the first index value; After using the time domain waveform file for differential discrete transformation, the standard deviation is used to measure the discrete degree to obtain the second index value; Combining the spectrum centroid and differential dispersion of the blade surface damage index, the monthly median is compared with the preset threshold to determine the ratio of the first index value and the second index value exceeding the preset threshold, and the maximum value of the three leaves is taken as the blade surface damage index; The process of determining the blade structure damage index includes: measuring from the theoretical change rate and relative change rate of the swing stiffness: The theoretical change rate of the shimmy stiffness is determined by the ratio of the peak position of the blade shimmy natural first-order frequency obtained through the amplitude spectrum diagram and the actual shimmy natural first-order frequency of the blade; The relative change rate of the shimmy stiffness is determined by the ratio of the peak position of the first-order frequency of the shimmy inherent of a blade obtained through the amplitude spectrum diagram to the average of the peak positions obtained by the remaining two blades; Combining the theoretical change rate and relative change rate of the shimmy stiffness, the monthly median is compared with the preset threshold to determine the proportion of the theoretical change rate and relative change rate of the shimmy stiffness exceeding the threshold, and the maximum value of the three blades is taken as the blade structure damage index; The process of determining the blade imbalance index includes: The flapping signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the flapping imbalance index; The shimmy signal of the time domain waveform file is used to calculate the RMS value near the 1p rotation frequency as the measurement value, and then the range of the three blades is calculated to determine the shimmy imbalance index; Combine the flapping imbalance index and the shimmy imbalance index, compare the monthly median with the preset threshold, and take the proportion of the flapping imbalance index and the shimmy imbalance index exceeding the preset threshold as the blade imbalance index; The process of determining blade icing monitoring indicators includes: Extract the blade temperature value in the working condition file, where the external environmental condition for blade icing is a temperature value below degrees; The process of determining the blade natural frequency monitoring index includes: the effective value of the first-order shimmy and the peak position of the first-order shimmy: Perform RMS statistics on the spectrum amplitude near the first-order natural frequency of the shimmy vibration, extract the peak position near the first-order natural frequency of the shimmy vibration, compare the monthly median with the threshold, and use the ratio of the effective value of the first-order shimmy vibration and the peak position of the first-order shimmy vibration exceeding the preset threshold as the blade natural frequency monitoring index; The process of determining the leaf data similarity index includes: The cosine similarity between the three blades of the amplitude spectrum is measured, and then the minimum similarity is statistically calculated to determine the frequency domain similarity; Calculate the time domain index, measure the cosine similarity between the three leaves of the time domain index sequence, and then calculate the minimum similarity to determine the time domain similarity; The data of the time domain waveform file is first processed by first-order difference, and then the time domain index value is counted, and then the cosine similarity between the three leaves of the sequence of the time domain index is measured, and then the minimum similarity is counted to determine the time domain difference similarity; The three types of indicators, namely frequency domain, time domain and time domain difference, are combined. The monthly median is compared with the preset threshold, and the proportion of the three types of indicators, namely frequency domain, time domain and time domain difference, exceeding the preset threshold is used as the leaf data similarity index.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the online monitoring method for wind turbine blades as claimed in claim 7 or 8 are implemented.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the online monitoring method for wind turbine blades as claimed in claim 7 or 8 are implemented.

Images