Wind turbine generator rotating speed signal analysis method and system based on flange plate bolt

Abstract

The invention discloses a wind turbine generator rotating speed signal analysis method and system based on a flange plate bolt, and the method comprises the following steps: collecting pulse basic information based on the flange plate bolt; judging whether the pulse width of the first pulse is abnormal or not; accumulating the original pulse information, identifying abnormal conditions in the original pulse information, and correcting pulse width change abnormity; counting and calculating the proportion of the abnormal pulses; pulse characteristics in the accumulated original pulse information are identified, N normal pulse signals are obtained according to the pulse characteristics, and the N normal pulse signals are used for replacing a single pulse period to calculate the initial impeller rotating speed; extracting pulses meeting a preset period, and adding the pulses into impeller rotating speed calculation; the occurrence probability of various abnormal features in the rotating speed is counted; according to the method, various abnormal signals are corrected, and normal bolt pulse signals are screened out, so that the accuracy of rotating speed calculation is improved, rotating speed abnormal fluctuation caused by interference factors is reduced, and more accurate data analysis is provided for a wind turbine generator.

Description

Technical Field

[0001] The present invention relates to the technical field of wind turbine speed signal analysis, and in particular to a method and system for analyzing wind turbine speed signals based on flange bolts. Background Art

[0002] The measurement and control of the rotor speed of wind turbines is related to the safety of the units. However, it is impossible to determine the abnormal state of the sensor to timely analyze and correct the original data, and thus it is impossible to obtain the true speed of the unit, resulting in unit failure and shutdown, or incorrect propeller opening, which increases the unit's extreme load, and even causes the unit to collapse. This problem has been common in the industry.

[0003] The current methods for measuring and calculating the rotor speed of wind turbines are:

[0004] 1) The speed signal of a direct-drive or semi-direct-drive wind turbine generator set can be calculated by the frequency of the generator stator voltage collected by the converter to calculate the generator speed, and then the wind turbine impeller speed is converted through the gearbox ratio. However, when the circuit breaker on the generator side is not closed, the converter cannot monitor the generator speed, and when the coupling slips, the generator impeller speed signal cannot be correctly judged.

[0005] 2) Use encoder to measure impeller speed:

[0006] A photoelectric encoder is installed at the tail of the slip ring, and the spindle speed data is collected and processed through a high sampling frequency counting module. Since the location of the encoder is easily affected by the vibration of the slip ring, the encoder speed jump may trigger the unit to shut down, which also makes the encoder easy to be damaged. In addition, the counting data processing method is limited by the configuration method of the purchased module, and the original data cannot be processed more flexibly.

[0007] 3) Calculate the impeller speed by collecting flange bolt pulses:

[0008] Install a proximity switch on the main shaft bolt or gearbox bolt, use the DI module to collect the proximity switch signal, and convert the collected pulse period into the impeller speed. Since the bolts used are polygonal, it is impossible to ensure that the width of each pulse detected is consistent during installation, and the distance between the proximity switch and the bolt, the measurement accuracy of the proximity switch, lightning weather, the position of the proximity switch during shutdown, the spacing between bolts, the main control cycle, whether the mounting bracket shakes, whether the bolt is concave, whether the bolt is missing, whether there are foreign objects between the bolts and other factors will affect the collected pulse signal. The use of filtering alone cannot completely suppress these interference factors that cause abnormal speed fluctuations, and the unit cannot predict and push the relevant status in advance to the relevant staff to understand and deal with potential risks, which can easily cause the whole machine control to fall into a high safety risk state because the real speed of the unit cannot be obtained. Summary of the invention

[0009] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a wind turbine speed signal analysis method and system based on flange bolts. By correcting various abnormal signals, normal bolt pulse signals are screened out, and the impeller speed is calculated, the accuracy of the speed calculation is improved, and the abnormal speed fluctuations caused by interference factors are reduced. The bolt status is automatically analyzed and pushed, providing a more accurate and efficient control solution and intelligent operation and maintenance management and data analysis method for the wind turbine.

[0010] The purpose of the present invention is achieved through the following technical solution: A method for analyzing the speed signal of a wind turbine generator set based on flange bolts, comprising the following steps:

[0011] S1. Collect basic pulse information based on flange bolts;

[0012] S2, determine whether there is an abnormality in the pulse width of the first pulse, if there is an abnormality, clear the pulse information in the buffer area and return to step S1, if there is no abnormality, execute step S3;

[0013] S3, accumulating the original pulse information, identifying the abnormal conditions in the original pulse information, and correcting the abnormal pulse width change; counting and calculating the proportion of abnormal pulses, if the proportion of abnormal pulses is greater than the preset threshold, after the abnormal bolt detection alarm of the wind turbine shutdown state is issued, the pulse information in the buffer area is cleared and the process returns to step S1; if the proportion of abnormal pulses is not greater than the preset threshold, step S4 is executed;

[0014] S4, identifying pulse features in the accumulated original pulse information, obtaining N normal pulse signals according to the pulse features, and using the N normal pulse signal cycles instead of a single pulse cycle to calculate the initial impeller speed;

[0015] S5, continuously detecting the pulse signal, extracting the pulses that meet the preset period, and adding them to the impeller speed calculation;

[0016] S6. Count the probability of various abnormal features appearing in the rotation speed, and send the relevant calculation and analysis results if it exceeds a preset probability threshold;

[0017] S7. When the calculated rotation speed of the new pulse cycle is lower than the preset rotation speed threshold, and the current pulse duration is too long, exceeding 1 minute, it is considered that the wind turbine is shut down and data is cleared.

[0018] Further, the step S1 comprises:

[0019] The basic pulse information includes the rising edge time, falling edge time, pulse period, high level duration and low level duration of all flange bolt pulses; and the pulse period, high level duration, low level duration and pulse width change information are collected and recorded.

[0020] Further, the step S2 comprises:

[0021] Determine the first pulse in the storage area. If the first pulse has extremely short pulse width, extremely short high level and extremely short low level, and the pulse width of adjacent pulses has irregular changes, it is considered to be an interference signal, and the pulse information in the cache area is cleared to return to step S1. If the first pulse is extremely long and the converted speed is low, it is considered that the wind turbine is completely shut down, and the pulse signals with the above characteristics are eliminated. Continue to screen the first pulse signal until a normal first pulse signal is screened out, and execute step S3.

[0022] Further, the step S3 comprises:

[0023] After obtaining the first pulse signal, the original pulse information is accumulated, and the proportion of abnormal pulse signals with short pulse period, short high level duration, short low level duration and pulse frequency change in the original pulse information is counted respectively. If the proportion of abnormal pulses is greater than the preset threshold or the proportion of abnormal pulses during the use of pulses is greater than the preset threshold, it is judged as an interference signal in the shutdown state, the pulse information in the cache area is cleared and returned to step S1, and the original pulses are accumulated again; if the proportion of abnormal pulses is not greater than the preset threshold, step S4 is executed.

[0024] Further, the step S4 comprises:

[0025] Identify and mark the pulse characteristics of the original pulse information in the storage area, and remove and correct the original pulse information according to the pulse characteristics. Obtain N normal pulse signals according to the pulse characteristics, and use the N normal pulse signals to replace a single pulse cycle to calculate the initial impeller speed n, in rpm, as shown in the following formula:

[0026] Impeller speed n = ((N / number of flange bolts in one circle) / sum of N pulse cycles)*60.

[0027] Further, the pulse characteristics include:

[0028] a) Normal feature 0: Regular waveform, and the rated pulse width and high and low level cycles are maintained within the preset range;

[0029] b) Abnormal feature 1: There are one or more high-level interference spikes between bolt pulses. The interference spikes can be combined with the previous bolt to form a pulse with normal pulse width. At the same time, the high-level cycle generated by the previous bolt pulse is normal.

[0030] c) Abnormal feature 2: The high level duration of the bolt pulse is short, and the low level duration in the entire pulse cycle accounts for a higher proportion than the normal cycle, but there is no obvious change in pulse width between the previous and next cycles;

[0031] d) Abnormal feature 3: Only two edge pulse signals of the bolt are detected. The pulse signal period of the first edge pulse signal is close to the high level period of the normal bolt, and the pulse period of the second edge pulse signal is close to the low level period of the normal bolt. The pulse period synthesized by the two edge pulses has no obvious change compared with the normal pulse period.

[0032] e) Abnormal feature 4: The bolt pulse signal is not detected, resulting in the pulse period being a multiple of the normal pulse period, and the high level duration has no significant change compared with the normal pulse period;

[0033] f) Abnormal feature 5: No bolt pulse gap is detected, the high level duration is twice or more than twice the normal pulse cycle duration, and the low level duration has no obvious change compared with the normal pulse cycle.

[0034] Further, the data removal and correction of the original pulse information according to the pulse characteristics includes:

[0035] Use normal pulses to filter out pulses that meet abnormal feature 1. If there are only one or more abnormal peaks, use the pulse width between the pulses before and after the abnormal peak as the Nth pulse width in the speed calculation pulse width, and remove the first pulse width cycle;

[0036] For pulses meeting abnormal characteristic 2, the pulse cycle is used normally;

[0037] For pulses that meet abnormal feature 3, the composite pulse width between the previous edge pulse signal and the next pulse signal after the next edge pulse signal is used as the Nth pulse width, and the first pulse width is removed;

[0038] For pulses that meet abnormal feature 4 and whose pulse width is a multiple of the normal pulse width, the pulse width is divided equally according to the abnormal multiple and used as the calculated pulse width, and the corresponding multiple pulse signals previously used for calculation are removed. If the abnormal pulse width is much larger than the normal pulse width, the abnormal pulse signal is not used;

[0039] For pulses that meet abnormal feature 5 and whose pulse width is twice or more than the normal pulse width, the pulse width is divided equally according to the abnormal multiple and used as the calculated pulse width, and the corresponding multiple pulse signals used for calculation are removed. If the abnormal pulse width far exceeds the normal pulse width, the abnormal pulse signal is not used.

[0040] Further, the step S5 comprises:

[0041] The probability of abnormal features appearing in each revolution within the preset time is classified and counted. If the probability reaches the set threshold, the analysis results will be pushed to the interface to prompt the staff to pay attention to the bolt status and make relevant processing; at the same time, the data of abnormal status is automatically stored for cause review and fault tracing.

[0042] A wind turbine speed signal analysis system based on flange bolts, used to implement the above-mentioned wind turbine speed signal analysis method based on flange bolts, comprising:

[0043] Proximity switch pulse signal acquisition module, used to collect pulse signals of flange bolts;

[0044] The DI module is used to convert the pulse signal collected by the proximity switch pulse signal collection module into a DI pulse signal;

[0045] The data processing module is used to process the DI pulse signal and analyze the pulse characteristics, obtain N normal DI pulse signals and historical data and store them;

[0046] The speed calculation module calculates the speed according to the processed N normal DI pulse signals;

[0047] Abnormal cause analysis module, used to analyze abnormal causes in DI pulse signals;

[0048] The abnormal data file generation module is used to generate an abnormal data file analysis report.

[0049] Furthermore, the data processing module includes:

[0050] Identify and mark the pulse characteristics of the original pulse information in the storage area, and remove and correct the data of the original pulse information according to the pulse characteristics, obtain N normal pulse signals according to the pulse characteristics, and use the N normal pulse signals to replace a single pulse cycle to calculate the initial impeller speed;

[0051] The pulse characteristics include:

[0052] a) Normal feature 0: Regular waveform, and the rated pulse width and high and low level cycles are maintained within the preset range;

[0053] b) Abnormal feature 1: There are one or more high-level interference spikes between bolt pulses. The interference spikes can be combined with the previous bolt to form a pulse with normal pulse width. At the same time, the high-level cycle generated by the previous bolt pulse is normal.

[0054] c) Abnormal feature 2: The high level duration of the bolt pulse is short, and the low level duration in the entire pulse cycle accounts for a higher proportion than the normal cycle, but there is no obvious change in pulse width between the previous and next cycles;

[0055] d) Abnormal feature 3: Only two edge pulse signals of the bolt are detected. The pulse signal period of the first edge pulse signal is close to the high level period of the normal bolt, and the pulse period of the second edge pulse signal is close to the low level period of the normal bolt. The pulse period synthesized by the two edge pulses has no obvious change compared with the normal pulse period.

[0056] e) Abnormal feature 4: The bolt pulse signal is not detected, resulting in the pulse period being a multiple of the normal pulse period, and the high level duration has no significant change compared with the normal pulse period;

[0057] f) Abnormal feature 5: No bolt pulse gap is detected, the high level duration is twice or more than the normal pulse cycle, and the low level duration has no significant change compared with the normal pulse cycle;

[0058] The data removal and correction of the original pulse information according to the pulse characteristics includes:

[0059] Use normal pulses to filter out pulses that meet abnormal feature 1. If there are only one or more abnormal peaks, use the pulse width between the pulses before and after the abnormal peak as the Nth pulse width in the speed calculation pulse width, and remove the first pulse width cycle;

[0060] For pulses meeting abnormal characteristic 2, the pulse cycle is used normally;

[0061] For pulses that meet abnormal feature 3, the composite pulse width between the previous edge pulse signal and the next pulse signal after the next edge pulse signal is used as the Nth pulse width, and the first pulse width is removed;

[0062] For pulses that meet abnormal feature 4 and whose pulse width is a multiple of the normal pulse width, the pulse width is divided equally according to the abnormal multiple and used as the calculated pulse width, and the corresponding multiple pulse signals previously used for calculation are removed. If the abnormal pulse width is much larger than the normal pulse width, the abnormal pulse signal is not used;

[0063] For pulses that meet abnormal feature 5 and whose pulse width is twice or more than the normal pulse width, the pulse width is divided equally according to the abnormal multiple and used as the calculated pulse width, and the corresponding multiple pulse signals used for calculation are removed. If the abnormal pulse width far exceeds the normal pulse width, the abnormal pulse signal is not used.

[0064] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0065] 1. The present invention obtains normal pulse signals to calculate the current impeller speed by cleaning and screening the original pulse data, which can better eliminate the influence of external interference factors, ensure the authenticity of the measured speed, avoid the unit entering the shutdown protection state due to abnormal interference and thus losing power generation; avoid the unit falling into safety risks due to the inability to measure the true speed.

[0066] 2. The present invention can automatically analyze the bolt status collected by the wind turbine, push the analyzed abnormal results to the interface, and save relevant data, which can help relevant staff better understand the status of the wind turbine and quickly locate the cause of the fault. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 The figure is a flow chart of the wind turbine speed signal analysis method based on flange bolts.

[0068] Figure 2 This is a schematic diagram of the interference signal waveform.

[0069] Figure 3 Schematic diagram of pulse characteristic waveform.

[0070] Figure 4 Schematic diagram of DI signal sampling and unprocessed speed waveform.

[0071] Figure 5 This is a schematic diagram of the speed waveform after processing.

[0072] Figure 6 This is a waveform diagram of the bolt missing speed processing.

[0073] Figure 7 This is the architecture diagram of the wind turbine speed signal analysis system based on flange bolts. DETAILED DESCRIPTION

[0074] The present invention will be further described below in conjunction with specific embodiments.

[0075] Example 1

[0076] See also Figure 1 As shown, the wind turbine speed signal analysis method based on flange bolts provided in this embodiment includes the following steps:

[0077] S1. Collect basic pulse information based on flange bolts, including:

[0078] The basic pulse information includes the rising edge time, falling edge time, pulse period, high level duration and low level duration of all flange bolt pulses; and the pulse period, high level duration, low level duration and pulse width change information are collected and recorded.

[0079] S2, judging whether there is an abnormality in the pulse width of the first pulse, if there is an abnormality, clearing the pulse information in the buffer area and returning to step S1, if there is no abnormality, executing step S3, including:

[0080] Determine the first pulse of the storage area. If the first pulse has extremely short pulse width, extremely short high level, extremely short low level pulse, and the pulse width of adjacent pulses changes irregularly, it is considered to be an interference signal. See Figure 2 As shown, clear the pulse information in the cache area and return to step S1; if the first pulse is extremely long and the converted speed is low, it is considered that the wind turbine is completely shut down, and the pulse signal with the above characteristics is eliminated, and the first pulse signal continues to be screened until a normal first pulse signal is screened out, and step S3 is executed.

[0081] S3, accumulating the original pulse information, identifying the abnormal situation in the original pulse information, and correcting the abnormal pulse width change; counting and calculating the proportion of abnormal pulses, if the proportion of abnormal pulses is greater than 1 / 4, after the abnormal bolt detection alarm of the wind turbine shutdown state is issued, the pulse information in the buffer area is cleared and returned to step S1; if the proportion of abnormal pulses is not greater than the preset threshold, executing step S4, including:

[0082] After obtaining the first pulse signal, the original pulse information is accumulated, and the proportion of abnormal pulse signals with short pulse period, short high level duration, short low level duration and pulse frequency change in the original pulse information is counted respectively. If the proportion of abnormal pulses is greater than 1 / 4 or the proportion of abnormal pulses during the use of pulses is greater than 1 / 4, it is judged as an interference signal in the shutdown state, the pulse information in the cache area is cleared and returned to step S1, and the original pulses are accumulated again; if the proportion of abnormal pulses is not greater than 1 / 4, step S4 is executed.

[0083] S4. Use multiple latest normal pulses to calculate the current speed. Since the bolts are polygonal, the edge-to-edge or corner-to-corner arrangement of the bolts after installation will affect the measured pulse period. In order to suppress the speed fluctuation of a single pulse caused by the installation position, the current speed is calculated using the N latest normal pulse signal periods instead of a single pulse period. Identify the pulse features in the accumulated original pulse information, obtain N normal pulse signals based on the pulse features, and use the N normal pulse signal periods instead of a single pulse period to calculate the initial impeller speed, including:

[0084] Identify and mark the pulse characteristics of the original pulse information in the storage area, and remove and correct the original pulse information according to the pulse characteristics. Obtain N normal pulse signals according to the pulse characteristics, and use the N normal pulse signals to replace a single pulse cycle to calculate the initial impeller speed n, in rpm, as shown in the following formula:

[0085] Impeller speed n = ((N / number of flange bolts in one circle) / N pulse cycle sum)*60; wherein, N pulse cycle sum is in seconds.

[0086] See also Figure 3 As shown, the pulse characteristics include:

[0087] a) Normal feature 0: Normal pulse, adjacent pulse widths are similar, and the proportions of high and low levels are similar. Considering the sampling accuracy and possible overspeed of the unit, the rated pulse width and high and low level cycles will not be too short;

[0088] b) Abnormal feature 1: There is a high-level interference spike between the bolt pulses. The interference spike can be combined with the previous bolt to form a normal pulse width pulse T5-T3. At the same time, the high-level cycle generated by the previous bolt pulse is normal. The position of the interference spike will affect the change rate of the pulse width of the front and rear bolts.

[0089] c) Abnormal feature 2: The high-level duration of the bolt pulse is short, and the low-level duration in the entire pulse cycle accounts for a higher proportion than the normal cycle. However, the entire pulse cycle T4-T3 is similar to the normal pulse cycle T5-T4, and there is no obvious pulse width change in the previous and next cycles. If combined with the previous and next cycles, the pulse width changes too quickly;

[0090] d) Abnormal feature 3: Only two edge pulse signals of the bolt are detected. The pulse signal period of the first edge pulse signal is close to the high level period of the normal bolt, and the pulse period of the second edge pulse signal is close to the low level period of the normal bolt. The pulse period T5-T3 synthesized by the two edge pulses has no obvious change compared with the normal pulse period.

[0091] e) Abnormal feature 4: The bolt pulse signal is not detected, resulting in the pulse period being a multiple of the normal pulse period, and the high level duration has no significant change compared with the normal pulse period;

[0092] f) Abnormal feature 5: No bolt pulse gap is detected, the high level duration is twice or more than twice the normal pulse cycle duration, and the low level duration has no obvious change compared with the normal pulse cycle.

[0093] The data removal and correction of the original pulse information according to the pulse characteristics includes:

[0094] Use normal pulses to filter out pulses that meet abnormal feature 1. If there are only one or more abnormal peaks, use T5-T3 as the rotation speed to calculate the Nth pulse width in the pulse width, and remove the first pulse width cycle;

[0095] For pulses meeting abnormal characteristic 2, the pulse cycle is used normally;

[0096] For pulses that meet abnormal feature 3, use T5-T3 as the Nth pulse width and remove the first pulse width;

[0097] For pulses that meet abnormal feature 4 and whose pulse width is a multiple of the normal pulse width, use half of T3-T2 as the N-1th and Nth pulse widths respectively, and remove the first and second pulse widths;

[0098] For pulses that meet abnormal feature 5 and whose pulse width is twice or more than the normal pulse width, use half of T4-T3 as the N-1th and Nth pulse widths respectively, and remove the first and second pulse widths.

[0099] The basic status of the newly added pulse information is normal, the previous pulse is a normal pulse, and the next pulse status information is normal.

[0100] S5, continuously detect the pulse signal, extract the pulse that meets the preset period, and add it to the impeller speed calculation, including:

[0101] The probability of abnormal features appearing in each revolution within the preset time is classified and counted. If the probability reaches the set threshold, the analysis results will be pushed to the interface to prompt the staff to pay attention to the bolt status and make relevant processing; at the same time, the data of abnormal status is automatically stored for cause review and fault tracing.

[0102] S6. Count the probability of various abnormal features appearing in the speed, and send the relevant calculation and analysis results if it exceeds the preset probability threshold of 1 / 3;

[0103] S7. When the calculated rotation speed of the new pulse cycle is lower than the preset rotation speed threshold and the current pulse duration is too long, exceeding 1 minute, it is considered that the wind turbine is shut down and the data is cleared.

[0104] See also Figures 4 to 5 As shown, the data processing of abnormal signal detection of bolts in shutdown state is as follows: when the wind turbine is in shutdown state, the abnormal interference signal DIsignal is generated due to the rapid jitter of the proximity switch on the edge of the bolt. The collected data is directly converted into the impeller speed grUnsortSpeed ​​without any processing. It can be seen that since the detection data is an interference signal with irregular period and extremely short pulse width, the impeller speed fluctuates significantly, and the calculated speed value far exceeds the rated speed of the unit, which easily causes the unit to enter a fault shutdown state; through the data processing method used in this embodiment, it is possible to identify, analyze and eliminate interference signals during unit startup detection, so that the speed will not generate high-speed signals due to abnormal jitter of the sensor. The processed speed grSpeedCal is 0, and the unit initial pulse detection abnormal alarm gbInitalDeteAbnBolt is triggered. The alarm is displayed on the interface and the relevant data will be cached in the form of files, which is convenient for relevant staff to understand the unit bolt detection status and cause tracing.

[0105] See also Figure 6 As shown, the data processing of bolt abnormality detection during normal operation of the unit is as follows: when a bolt signal is lost during the operation of the unit and no data processing is performed during the speed calculation, it will be detected that the pulse width has doubled, which will cause the speed grUnsortSpeed ​​to drop by half. However, in this patent, the pulse width change and pulse width length are automatically identified during data processing, and the speed grSpeedCal is calculated after the missing pulse width is automatically filled when it is determined to be a bolt. It can be seen that the speed fluctuates smoothly and the calculated value is closer to the actual speed.

[0106] Example 2

[0107] See also Figure 7 As shown, the wind turbine speed signal analysis system based on flange bolts provided in this embodiment is used to implement the wind turbine speed signal analysis method based on flange bolts described in Example 1, including:

[0108] Proximity switch pulse signal acquisition module, used to collect pulse signals of flange bolts;

[0109] The DI module is used to convert the pulse signal collected by the proximity switch pulse signal collection module into a DI pulse signal;

[0110] The data processing module is used to process the DI pulse signal and analyze the pulse characteristics, obtain N normal DI pulse signals and historical data and store them, including:

[0111] Identify and mark the pulse characteristics of the original pulse information in the storage area, and remove and correct the data of the original pulse information according to the pulse characteristics, obtain N normal pulse signals according to the pulse characteristics, and use the N normal pulse signals to replace a single pulse cycle to calculate the initial impeller speed;

[0112] The pulse characteristics include:

[0113] a) Normal feature 0: Regular waveform, and the rated pulse width and high and low level cycles are maintained within the preset range;

[0114] b) Abnormal feature 1: There are one or more high-level interference spikes between bolt pulses. The interference spikes can be combined with the previous bolt to form a pulse with normal pulse width. At the same time, the high-level cycle generated by the previous bolt pulse is normal.

[0115] c) Abnormal feature 2: The high level duration of the bolt pulse is short, and the low level duration in the entire pulse cycle accounts for a higher proportion than the normal cycle, but there is no obvious change in pulse width between the previous and next cycles;

[0116] d) Abnormal feature 3: Only two edge pulse signals of the bolt are detected. The pulse signal period of the first edge pulse signal is close to the high level period of the normal bolt, and the pulse period of the second edge pulse signal is close to the low level period of the normal bolt. The pulse period synthesized by the two edge pulses has no obvious change compared with the normal pulse period.

[0117] e) Abnormal feature 4: The bolt pulse signal is not detected, resulting in the pulse period being a multiple of the normal pulse period, and the high level duration has no significant change compared with the normal pulse period;

[0118] f) Abnormal feature 5: No bolt pulse gap is detected, the high level duration is twice or more than the normal pulse cycle, and the low level duration has no significant change compared with the normal pulse cycle;

[0119] The data removal and correction of the original pulse information according to the pulse characteristics includes:

[0120] Use normal pulses to filter out pulses that meet abnormal feature 1. If there are only one or more abnormal peaks, use the pulse width between the pulses before and after the abnormal peak as the Nth pulse width in the speed calculation pulse width, and remove the first pulse width cycle;

[0121] For pulses meeting abnormal characteristic 2, the pulse cycle is used normally;

[0122] For pulses that meet abnormal feature 3, the pulse width between the previous edge pulse signal and the next pulse signal after the next edge pulse signal is used as the Nth pulse width, and the first pulse width is removed;

[0123] For pulses that meet abnormal feature 4 and whose pulse width is a multiple of the normal pulse width, the pulse width is divided equally according to the abnormal multiple and used as the calculated pulse width, and the corresponding multiple pulse signals previously used for calculation are removed. If the abnormal pulse width is much larger than the normal pulse width, the abnormal pulse signal is not used;

[0124] For pulses that meet abnormal feature 5 and whose pulse width is twice or more than the normal pulse width, the pulse width is divided equally according to the abnormal multiple and used as the calculated pulse width, and the corresponding multiple pulse signals used for calculation are removed. If the abnormal pulse width far exceeds the normal pulse width, the abnormal pulse signal is not used.

[0125] The speed calculation module calculates the speed according to the processed N normal DI pulse signals;

[0126] Abnormal cause analysis module, used to analyze abnormal causes in DI pulse signals;

[0127] The abnormal data file generation module is used to generate an abnormal data file analysis report.

[0128] Example 3

[0129] This embodiment discloses a non-transitory computer-readable medium storing instructions. When the instructions are executed by a processor, the steps of the wind turbine speed signal analysis method based on flange bolts according to Embodiment 1 are performed.

[0130] The non-temporary computer-readable medium in this embodiment may be a disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), a USB flash drive, a mobile hard disk, or other media.

[0131] Example 4

[0132] This embodiment discloses a computing device, including a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the method for analyzing a wind turbine speed signal based on flange bolts described in Example 1 is implemented.

[0133] The computing device described in this embodiment may be a desktop computer, a laptop computer, a smart phone, a PDA handheld terminal, a tablet computer, a programmable logic controller (PLC), or other terminal devices with a processor function.

[0134] The embodiments described above are only preferred embodiments of the present invention and are not intended to limit the scope of implementation of the present invention. Therefore, all changes made according to the shape and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A wind turbine speed signal analysis method based on flange bolts, It is characterized in that The following steps are involved: S1. Collect basic pulse information based on flange bolts; S2, determine whether there is an abnormality in the pulse width of the first pulse, if there is an abnormality, clear the pulse information in the buffer area and return to step S1, if there is no abnormality, execute step S3; S3, accumulating original pulse information, identifying abnormalities in the original pulse information, and correcting abnormal pulse width changes; The proportion of abnormal pulses is counted and calculated. If the proportion of abnormal pulses is greater than a preset threshold, an abnormal bolt detection alarm is issued in the shutdown state of the wind turbine, and the pulse information in the buffer area is cleared and the process returns to step S1. If the proportion of abnormal pulses is not greater than the preset threshold, step S4 is executed. S4, identifying pulse features in the accumulated original pulse information, obtaining N normal pulse signals according to the pulse features, and using the N normal pulse signal cycles instead of a single pulse cycle to calculate the initial impeller speed; S5, continuously detecting the pulse signal, extracting the pulses that meet the preset period, and adding them to the impeller speed calculation; S6. Count the probability of various abnormal features appearing in the rotation speed, and send the relevant calculation and analysis results if it exceeds a preset probability threshold; S7. When the calculated rotation speed of the new pulse cycle is lower than the preset rotation speed threshold, and the current pulse duration is too long, exceeding 1 minute, it is considered that the wind turbine is shut down and data is cleared.

2. A wind turbine speed signal analysis method based on flange bolts according to claim 1, It is characterized in that The step S1 comprises: The basic pulse information includes the rising edge time, falling edge time, pulse period, high level duration and low level duration of all flange bolt pulses; and the pulse period, high level duration, low level duration and pulse width change information are collected and recorded.

3. According to claim 1, a method for analyzing a wind turbine speed signal based on flange bolts, It is characterized in that The step S2 comprises: Determine the first pulse in the storage area. If the first pulse has extremely short pulse width, extremely short high level and extremely short low level, and the pulse width of adjacent pulses has irregular changes, it is considered to be an interference signal, and the pulse information in the cache area is cleared to return to step S1. If the first pulse is extremely long and the converted speed is low, it is considered that the wind turbine is completely shut down, and the pulse signals with the above characteristics are eliminated. Continue to screen the first pulse signal until a normal first pulse signal is screened out, and execute step S3.

4. According to claim 1, a method for analyzing a wind turbine speed signal based on flange bolts, It is characterized in that The step S3 comprises: After obtaining the first pulse signal, the original pulse information is accumulated, and the proportion of abnormal pulse signals with short pulse period, short high level duration, short low level duration and pulse frequency change in the original pulse information is counted respectively. If the proportion of abnormal pulses is greater than the preset threshold or the proportion of abnormal pulses during the use of pulses is greater than the preset threshold, it is judged as an interference signal in the shutdown state, the pulse information in the cache area is cleared and returned to step S1, and the original pulses are accumulated again; if the proportion of abnormal pulses is not greater than the preset threshold, step S4 is executed.

5. The method for analyzing the speed signal of a wind turbine generator system based on flange bolts according to claim 1, It is characterized in that The step S4 comprises: Identify and mark the pulse characteristics of the original pulse information in the storage area, and remove and correct the original pulse information according to the pulse characteristics. Obtain N normal pulse signals according to the pulse characteristics, and use the N normal pulse signals to replace a single pulse cycle to calculate the initial impeller speed n, in rpm, as shown in the following formula: Impeller speed n = ((N / number of flange bolts in one circle) / N pulse cycles sum)*60; Where N is the pulse period and the unit is seconds.

6. A method for analyzing a wind turbine speed signal based on flange bolts according to claim 5, It is characterized in that The pulse characteristics include: a) Normal feature 0: Regular waveform, and the rated pulse width and high and low level cycles are maintained within the preset range; b) Abnormal feature 1: There are one or more high-level interference spikes between bolt pulses. The interference spikes can be combined with the previous bolt to form a pulse with normal pulse width. At the same time, the high-level cycle generated by the previous bolt pulse is normal. c) Abnormal feature 2: The high level duration of the bolt pulse is short, and the low level duration in the entire pulse cycle accounts for a higher proportion than the normal cycle, but there is no obvious pulse width change in the previous and next cycles; d) Abnormal feature 3: Only two edge pulse signals of the bolt are detected. The pulse signal period of the first edge pulse signal is close to the high level period of the normal bolt, and the pulse period of the second edge pulse signal is close to the low level period of the normal bolt. The pulse period synthesized by the two edge pulses has no obvious change compared with the normal pulse period. e) Abnormal feature 4: The bolt pulse signal is not detected, resulting in the pulse period being a multiple of the normal pulse period, and the high level duration has no significant change compared with the normal pulse period; f) Abnormal feature 5: No bolt pulse gap is detected, the high level duration is twice or more than twice the normal pulse cycle duration, and the low level duration has no obvious change compared with the normal pulse cycle.

7. A method for analyzing a wind turbine speed signal based on flange bolts according to claim 6, It is characterized in that The data removal and correction of the original pulse information according to the pulse characteristics includes: Use normal pulses to filter out pulses that meet abnormal feature 1. If there are only one or more abnormal peaks, use the pulse width between the pulses before and after the abnormal peak as the Nth pulse width in the speed calculation pulse width, and remove the first pulse width cycle; For pulses meeting abnormal characteristic 2, the pulse cycle is used normally; For pulses that meet abnormal feature 3, the composite pulse width between the previous edge pulse signal and the next pulse signal after the next edge pulse signal is used as the Nth pulse width, and the first pulse width is removed; For pulses that meet abnormal feature 4 and whose pulse width is a multiple of the normal pulse width, the pulse width is divided equally according to the abnormal multiple and used as the calculated pulse width, and the corresponding multiple pulse signals previously used for calculation are removed. If the abnormal pulse width is much larger than the normal pulse width, the abnormal pulse signal is not used; For pulses that meet abnormal feature 5 and whose pulse width is twice or more than the normal pulse width, the pulse width is divided equally according to the abnormal multiple and used as the calculated pulse width, and the corresponding multiple pulse signals used for calculation are removed. If the abnormal pulse width far exceeds the normal pulse width, the abnormal pulse signal is not used.

8. The method for analyzing the wind turbine speed signal based on flange bolts according to claim 1, It is characterized in that The step S5 comprises: The probability of abnormal features appearing in each revolution within the preset time is classified and counted. If the probability reaches the set threshold, the analysis results will be pushed to the interface to prompt the staff to pay attention to the bolt status and make relevant processing; at the same time, the data of abnormal status is automatically stored for cause review and fault tracing.

9. A wind turbine speed signal analysis system based on flange bolts, It is characterized in that The method for analyzing the speed signal of a wind turbine generator system based on flange bolts according to any one of claims 1 to 8 comprises: Proximity switch pulse signal acquisition module, used to collect pulse signals of flange bolts; The DI module is used to convert the pulse signal collected by the proximity switch pulse signal collection module into a DI pulse signal; The data processing module is used to process the DI pulse signal and analyze the pulse characteristics, obtain N normal DI pulse signals and historical data and store them; The speed calculation module calculates the speed according to the processed N normal DI pulse signals; Abnormal cause analysis module, used to analyze abnormal causes in DI pulse signals; The abnormal data file generation module is used to generate an abnormal data file analysis report.

10. A wind turbine speed signal analysis system based on flange bolts according to claim 9, It is characterized in that The data processing module comprises: Identify and mark the pulse characteristics of the original pulse information in the storage area, and remove and correct the data of the original pulse information according to the pulse characteristics, obtain N normal pulse signals according to the pulse characteristics, and use the N normal pulse signals to replace a single pulse cycle to calculate the initial impeller speed; The pulse characteristics include: a) Normal feature 0: Regular waveform, and the rated pulse width and high and low level cycles are maintained within the preset range; b) Abnormal feature 1: There are one or more high-level interference spikes between bolt pulses. The interference spikes can be combined with the previous bolt to form a pulse with normal pulse width. At the same time, the high-level cycle generated by the previous bolt pulse is normal. c) Abnormal feature 2: The high level duration of the bolt pulse is short, and the low level duration in the entire pulse cycle accounts for a higher proportion than the normal cycle, but there is no obvious pulse width change in the previous and next cycles; d) Abnormal feature 3: Only two edge pulse signals of the bolt are detected. The pulse signal period of the first edge pulse signal is close to the high level period of the normal bolt, and the pulse period of the second edge pulse signal is close to the low level period of the normal bolt. The pulse period synthesized by the two edge pulses has no obvious change compared with the normal pulse period. e) Abnormal feature 4: The bolt pulse signal is not detected, resulting in the pulse period being a multiple of the normal pulse period, and the high level duration has no significant change compared with the normal pulse period; f) Abnormal feature 5: No bolt pulse gap is detected, the high level duration is twice or more than the normal pulse cycle, and the low level duration has no significant change compared with the normal pulse cycle; The data removal and correction of the original pulse information according to the pulse characteristics includes: Use normal pulses to filter out pulses that meet abnormal feature 1. If there are only one or more abnormal peaks, use the pulse width between the pulses before and after the abnormal peak as the Nth pulse width in the speed calculation pulse width, and remove the first pulse width cycle; For pulses meeting abnormal characteristic 2, the pulse cycle is used normally; For pulses that meet abnormal feature 3, the composite pulse width between the previous edge pulse signal and the next pulse signal after the next edge pulse signal is used as the Nth pulse width, and the first pulse width is removed; For pulses that meet abnormal feature 4 and whose pulse width is a multiple of the normal pulse width, the pulse width is divided equally according to the abnormal multiple and used as the calculated pulse width, and the corresponding multiple pulse signals previously used for calculation are removed. If the abnormal pulse width is much larger than the normal pulse width, the abnormal pulse signal is not used; For pulses that meet abnormal feature 5 and whose pulse width is twice or more than the normal pulse width, the pulse width is divided equally according to the abnormal multiple and used as the calculated pulse width, and the corresponding multiple pulse signals used for calculation are removed. If the abnormal pulse width far exceeds the normal pulse width, the abnormal pulse signal is not used.

Images