A bolt loosening monitoring method and monitoring device

Abstract

A bolt loosening monitoring method and monitoring device, a bolt monitoring signal collection and processing method is proposed, a bolt loosening coefficient calculation method is obtained combined with the method, and a bolt pre-loosening early warning method is obtained; and a bolt dynamic strain average growth coefficient calculation method is proposed; the above data processing is applied, and the required signals are collected to obtain a bolt loosening monitoring device; solve the problem of the current fan flange bolt loosening monitoring method and device, can reduce the occurrence of safety accidents caused by fan flange bolt loosening, and scientifically guide the daily operation and maintenance of the fan to further reduce the daily maintenance cost.

Description

Technical Field

[0001] This invention relates to the field of structural safety monitoring methods and devices, specifically to a bolt loosening monitoring method and device. Background Technology

[0002] Loosening of flange bolts is a common fault during the operation of wind turbines. Currently, there are no mature monitoring methods and devices. The existing bolt loosening monitoring methods and devices are mainly: (1) using strain gauges to monitor the stress of bolts; (2) using ultrasonic waves to detect the bolt length or using angle sensors to detect the relative angle change between the bolt and nut. Each of these three methods has its own advantages and disadvantages. In (1), strain gauges can very sensitively sense changes in bolt strain and stress, but their output values ​​are also sensitive to external loads. Therefore, the reliability of the monitoring results of this method is poor. In (2), ultrasonic detection is greatly affected by ambient temperature. When used for long-term monitoring, temperature drift compensation devices are required, which leads to complicated installation process and high monitoring costs, thus limiting its application range.

[0003] Moreover, after the acquired signal is subjected to surrounding electromagnetic interference, the acquired signal is mixed with this random disturbance that cannot be eliminated, resulting in data distortion and ineffective utilization; therefore, there is currently a lack of effective technical measures to eliminate interference and remove its effects on the acquired signal.

[0004] Currently, the common method for detecting loose bolts is to use sensors to monitor their deformation. Traditionally, this method relies on the magnitude of the deformation, i.e., the scale of change in deformation, and lacks an effective and predictive method for calculating bolt loosening. Furthermore, relying on sensor monitoring only addresses the result of deformation that has already occurred. Relying on deformation-based replacement and maintenance presents two problems: if the deformation is set too high, the bolt may already be damaged, creating a safety hazard, providing only a post-incident warning and failing to effectively prevent potential safety risks to equipment operation; if the deformation is set too low, replacement and maintenance based on this will introduce random errors, and the bolt may still have a service life margin, preventing it from fully utilizing its intended lifespan. Therefore, this traditional method of measuring strain using a single sensor results in either overly frequent or premature replacement, or premature replacement or replacement after bolt damage, meaning that the operational risks of potential hazards cannot be avoided.

[0005] Therefore, there is currently a lack of effective early warning methods for the structural risks of bolt loosening, as well as a lack of devices that can effectively monitor bolt loosening. Summary of the Invention

[0006] In view of this, and in view of the shortcomings of the existing technology, the purpose of this invention is to provide a novel bolt loosening monitoring method: by deeply integrating the bolt strain monitoring data and the nut relative rotation data, a bolt loosening probability assessment and loosening early warning model, and a bolt loosening impact assessment and loosening alarm model are established, thereby realizing trend early warning before the wind turbine flange bolts loosen and alarm of the degree of danger after loosening; based on this method, a new bolt loosening monitoring device is designed.

[0007] This application aims to address one of the problems in the background art.

[0008] The technical solution adopted in this invention is as follows: In order to achieve the above-mentioned objectives and other related objectives, this invention provides a bolt monitoring signal acquisition and processing method, which couples the nut rotation angle and dynamic strain change value acquired within a time period t, extracts the values ​​greater than or equal to 0, and removes the values ​​less than 0; the coupling operation is: the nut rotation angle and dynamic strain change value acquired within a time period t are multiplied one-to-one.

[0009] The technical solution provided in this application also has the following technical features:

[0010] Preferably, at least one set of A and B is collected within time period t;

[0011] A represents a sequence of n numerical changes in the dynamic strain of the bolt relative to its initial value: ;

[0012] B represents a sequence of n numerical values ​​showing the change in the nut's rotation angle relative to its initial value. ;

[0013] Sequence A multiplied by sequence B equals sequence C: ;

[0014] n≥z≥1, and z is an integer, satisfying the condition in the C sequence. Values ​​greater than or equal to 0:

[0015] extract To obtain the data sequence ;

[0016] extract To obtain the data sequence .

[0017] Preferred, extraction ,extract The order is consistent, z ranges from 1 to n, and the filter matches The corresponding condition for being greater than or equal to 0 and .

[0018] Preferred, data sequence , The average strain of the bolt during time interval t is calculated by performing mean value calculations, which is the characteristic value of that time interval t. , ;

[0019] From the first t-period to the m-th t-period, a sequence containing m feature values ​​was obtained through continuous monitoring:

[0020] Dynamic strain mean sequence, ;

[0021] The mean sequence of nut rotation angles .

[0022] Applying the above-mentioned bolt monitoring signal acquisition and processing method, a method for calculating the bolt loosening coefficient is obtained. Within K time intervals (mt), a sample sequence is obtained:

[0023] , ;

[0024] A continuously increasing sequence: each value in the sequence is not less than its next adjacent value.

[0025] That is, each value in the sequence satisfies: ;

[0026] in, The number of consecutively increasing sequences in the sequence is denoted as S;

[0027] in, The number of consecutively increasing sequences in the sequence is denoted as T;

[0028] Probability calculations for assessing growth trends are performed on the eigenvalue sequences respectively:

[0029] The probability that the dynamic strain mean sequence shows an increasing trend is: P(A) = S / K;

[0030] The probability that the mean sequence of nut rotation angles shows an increasing trend is: P(B) = T / K;

[0031] Bolt loosening coefficient P(C) = βP(A) + (1-β)P(B); 0≤β≤1, where β is a correction coefficient and is related to the specifications and materials of the bolt and nut;

[0032] Substituting P(A) and P(B), P(C) = (βS + T - βT) / K.

[0033] Preferably, between the K mt time periods, each adjacent mt time period is either continuous, equally spaced, or non-equally spaced.

[0034] Preferably, t / n is 1 second.

[0035] Preferably, the time period t is one minute and n is 60.

[0036] Preferably, there are K MT periods within one month.

[0037] By applying the above-mentioned method for calculating the bolt loosening coefficient, a bolt pre-loosening warning method is obtained: if P(C) < 0.5, the bolt is safe and no warning is issued; if P(C) ≥ 0.5, a bolt loosening warning is issued.

[0038] A method for calculating the growth coefficient of the mean dynamic strain of bolts involves performing mean growth screening on a dynamic strain mean value sequence, or one or more dynamic strain mean value sequences connected in a time sequence. Let the dynamic strain mean value sequence be... ;

[0039] In d, among adjacent values, h ranges from 1 to p, which does not conform to the condition. Remove Obtain the sequence ;

[0040] For sequence Repeatedly perform mean-growth filtering until the sequence... All means in the range conform to ;

[0041] sequence If the mean number of values ​​is r, then the growth coefficient is: P(d) = r / p.

[0042] This technical solution also has the following technical features:

[0043] Preferably, a bolt loosening early warning method is obtained by applying a bolt dynamic strain mean growth coefficient calculation method to obtain P(d), P(C)≥0.5, and: the absolute value of the maximum or minimum value of the nut rotation angle change is greater than 1°, and the dynamic strain mean growth coefficient P(d) is greater than 0.5; and a bolt loosening early warning is issued.

[0044] A bolt loosening monitoring device includes a bolt dynamic strain sensing module, a nut rotation angle sensing module, and a data acquisition and analysis module. The data acquisition and analysis module is connected to the bolt dynamic strain sensing module and the nut rotation angle sensing module.

[0045] The bolt dynamic strain sensing module is used to monitor the dynamic strain value of flange bolts;

[0046] The nut rotation angle sensing module is used to monitor the rotation angle value of the nut relative to the bolt;

[0047] The data acquisition and analysis module is used to acquire signals from the bolt dynamic strain sensing module and the nut rotation angle sensing module.

[0048] This technical solution also has the following technical features:

[0049] Preferably, the bolt dynamic strain sensing module is a full-bridge resistive dynamic strain gauge, which is welded to the bolt at both ends; the nut rotation angle sensing module is fastened to the nut.

[0050] Preferably, the data acquisition and analysis module extracts the feature values ​​of each monitoring parameter online, performs coupling operation on the acquired signal and calculates the mean value to obtain the mean value sequence, and obtains T, S, and K; the bolt loosening probability assessment and early warning module is used to calculate P(C).

[0051] Preferably, the bolt loosening monitoring device is used for monitoring flange bolt loosening, and the bolt dynamic strain sensing module and the nut rotation angle sensing module are connected to the data acquisition and analysis module through a sampling switch of the same frequency.

[0052] Preferably, the sampling switch with the same frequency is a synchronous sampler with equal period.

[0053] Preferably, the bolt loosening probability assessment and early warning module is used to assess the probability and trend of bolts loosening, and to issue loosening early warning information for bolts with a high probability of loosening.

[0054] Preferably, a bolt loosening degree and safety alarm module is used to assess the impact of bolt loosening and issue a loosening hazard alarm.

[0055] The present invention has the following beneficial effects:

[0056] 1. This invention addresses the problem of loose bolts, especially loose bolts on wind turbine flanges, but there is no mature monitoring method or device available. This invention creatively proposes a low-cost and reliable method for monitoring loose bolts on wind turbine flanges and designs a corresponding device.

[0057] 2. This invention employs a bolt monitoring signal acquisition and processing method. Through coupling operations, electromagnetic interference signal values ​​are eliminated, removing interference and random error values ​​to obtain effective and accurate acquisition signals, providing technical support for accurate prediction of bolt loosening. The principle is as follows: when a bolt is stretched, the dynamic strain value increases, and vice versa; when a bolt is stretched, the angle between the nut and the bolt increases, and vice versa. Therefore, before the bolt prestress is completely lost, the output values ​​of the dynamic strain gauge and angle sensor have consistent directions of change. During the coupling operation, values ​​greater than or equal to 0 represent strain and angle changes in the same direction, indicating consistency and are considered valid data; values ​​less than 0 represent strain and angle changes in a different direction, indicating inconsistency and are considered interference data.

[0058] 3. After eliminating interference signals, this invention utilizes coupling operations to obtain the acquired signals and performs averaging to obtain an average sequence, providing data support for the subsequent calculation of the bolt loosening coefficient. This calculation method, based on multiple average sequences of dynamic strain and nut rotation angles acquired from the bolt, statistically analyzes the increasing sequences to determine the continuous changes in the bolt. Based on the statistical analysis of the changes in the acquired average sequences, the bolt loosening coefficient is obtained. Specifically, the bolt exhibits continuous growth across multiple time intervals (mt), and within each individual mt time interval of the increasing sequence, the average dynamic strain also continuously increases. This invention provides an effective method for estimating the bolt loosening coefficient, offering timely signal processing to obtain a safety assessment of bolt loosening, providing a reliable basis for inspection and maintenance. This invention addresses the shortcomings of current wind turbine flange bolt loosening monitoring methods and devices, reducing safety accidents caused by wind turbine flange bolt loosening and scientifically guiding daily wind turbine operation and maintenance, thereby reducing daily maintenance costs.

[0059] 4. This invention also provides an application processing method that utilizes the maximum value of the acquired signal. When the absolute value of the maximum or minimum value of the nut rotation angle change is greater than 1°, and the growth coefficient P(d) of the average dynamic strain is greater than 0.5, this abrupt change is used to estimate the risk of bolt loosening. This serves as an indicator signal for bolt loosening. It is very effective in eliminating interference or compensating for insufficient average value sequences collected in the initial stage, such as k being less than 30 and t being one minute. This is because in the early stages of bolt use and monitoring, the continuous increase of P(A), P(B), and P(C), coupled with the small number of data samples, leads to a large bolt loosening coefficient, thus estimating a high risk of bolt loosening. Attached Figure Description

[0060] Figure 1 This is a schematic diagram of a bolt loosening monitoring device according to the present invention;

[0061] Figure 2 This is a schematic flowchart of a bolt loosening monitoring method according to the present invention;

[0062] Figure 3 This is a schematic diagram of the installation state of the dynamic strain gauge of the bolt loosening monitoring device according to the present invention;

[0063] Figure 4 This is a schematic diagram of an angle sensor for a bolt loosening monitoring device according to the present invention;

[0064] The markings in the diagram are explained as follows:

[0065] 1. Bolts

[0066] 2. Dynamic strain gauge

[0067] 3. Angle sensor

[0068] 4. Nuts. Detailed Implementation

[0069] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings. These embodiments are for illustrative purposes only and are not intended to limit the scope of the invention.

[0070] In the description of this invention, it should be noted that the terms "center," "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0071] In the description of this invention, it should be noted that, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0072] Furthermore, in the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0073] A bolt monitoring signal acquisition and processing method involves coupling the nut rotation angle and dynamic strain change value acquired within a time period t, extracting values ​​greater than or equal to 0, and removing values ​​less than 0. The coupling operation is as follows: the nut rotation angle and dynamic strain change value acquired within a time period t are multiplied one-to-one.

[0074] This invention has the following characteristics: by utilizing coupling operation, electromagnetic interference and abnormal deformation of both are eliminated; by utilizing the synergy of such a changing trend, the dynamic strain value increases when the bolt is stretched, and decreases when stretched.

[0075] When the bolt is stretched, the angle between the nut and the bolt increases, and vice versa.

[0076] Therefore, before the bolt prestress is completely lost, the output values ​​of the dynamic strain gauge and angle sensor have a consistent direction of change.

[0077] Specifically, at least one set of A and B is collected within the time period t; the time period t is determined based on the application scenario and the sensor sampling frequency, and can be one minute with one sample per second; or the time period t can be one hour with one sample per second.

[0078] A represents a sequence of n numerical changes in the dynamic strain of the bolt relative to its initial value: ;

[0079] B represents a sequence of n numerical values ​​showing the change in the nut's rotation angle relative to its initial value. ;

[0080] Sequence A multiplied by sequence B equals sequence C: ;

[0081] n≥z≥1, and z is an integer, satisfying the condition in the C sequence. Values ​​greater than or equal to 0:

[0082] extract To obtain the data sequence ;

[0083] extract To obtain the data sequence We obtained the mean series after eliminating interference, which provides a reliable data foundation for subsequent analysis.

[0084] Specifically, extract ,extract The order is consistent, z ranges from 1 to n, and the filter matches The corresponding condition for being greater than or equal to 0 and .

[0085] Specifically, data sequence , The average strain of the bolt during time interval t is calculated by performing mean value calculations, which is the characteristic value of that time interval t. , ;

[0086] From the first t-period to the m-th t-period, a sequence containing m feature values ​​was obtained through continuous monitoring:

[0087] Dynamic strain mean sequence, ;

[0088] The mean sequence of nut rotation angles .

[0089] Applying the above-mentioned bolt monitoring signal acquisition and processing method, a method for calculating the bolt loosening coefficient is obtained. Within K time intervals (mt), a sample sequence is obtained:

[0090] , ;

[0091] A continuously increasing sequence: each value in the sequence is not less than its next adjacent value.

[0092] That is, each value in the sequence satisfies: ;

[0093] in, The number of consecutively increasing sequences in the sequence is denoted as S;

[0094] in, The number of consecutively increasing sequences in the sequence is denoted as T;

[0095] Probability calculations for assessing growth trends are performed on the eigenvalue sequences respectively:

[0096] The probability that the mean dynamic strain sequence shows an increasing trend is: P(A) = S / K;

[0097] The probability that the mean rotation angle sequence of the nut shows an increasing trend is: P(B) = T / K;

[0098] Bolt loosening coefficient P(C) = βP(A) + (1-β)P(B); 0≤β≤1, where β is a correction coefficient and is related to the specifications and materials of the bolt and nut;

[0099] Substituting P(A) and P(B), P(C) = (βS + T - βT) / K;

[0100] Depending on the application scenario, an initial β coefficient can be established, with β=0.5. Then, simulation or experimentation can be performed, and β can be corrected to make it close to the simulation result. When the error is less than 5%, the β value of the iterative correction is the correction coefficient for the bolt specification.

[0101] Specifically, the loosening coefficient of a certain bolt on the fan flange was tested and recorded after it became loose.

[0102] The condition of the bolts was observed and verified, and the following data and observation results were obtained:

[0103] K=100;

[0104]

[0105] Substitute the initial values ​​of β into 0.4, 0.5, and 0.6, calculate P(C), check whether it matches the observation results, and adjust the direction of β.

[0106]

[0107] Therefore, we further adjusted β to 0.7, 0.8, and 0.9 to check whether it matched the observation results;

[0108]

[0109] Therefore, β is between 0.7 and 0.8. By obtaining more samples, β can be further corrected. Therefore, the β observed and calibrated in this study is 0.75. This correction factor can be used to estimate subsequent bolt loosening. Substituting β=0.75, P(C)=(0.75S+0.25T) / K.

[0110] Specifically, among the K mt time periods, each adjacent mt time period can be continuous, equally spaced, or not equally spaced.

[0111] By applying the above-mentioned method for calculating the bolt loosening coefficient, a bolt pre-loosening warning method is obtained: if P(C) < 0.5, the bolt is safe and no warning is issued; if P(C) ≥ 0.5, a bolt loosening warning is issued.

[0112] A method for calculating the growth coefficient of the mean dynamic strain of bolts involves performing mean growth screening on a dynamic strain mean value sequence, or one or more dynamic strain mean value sequences connected in a time sequence. Let the dynamic strain mean value sequence be... ;

[0113] In d, among adjacent values, h ranges from 1 to p, which does not conform to the condition. Remove Obtain the sequence ;

[0114] For sequence Repeatedly perform mean-growth filtering until the sequence... All means in the range conform to ;

[0115] sequence If the mean number of values ​​is r, then the growth coefficient is: P(d) = r / p. This growth coefficient can be used to estimate the rate of change of dynamic strain and its continuous growth. Since dynamic strain is continuously changing, that is, expansion and contraction, and only when the strain is larger each time can the bolt be damaged, including fatigue failure and one-time overload failure. Therefore, this growth coefficient can be used to estimate whether the continuous strain of the bolt is within a safe range.

[0116] Specifically, a bolt loosening early warning method is obtained. A growth coefficient calculation method for the average dynamic strain of the bolt is applied to obtain P(d), P(C)≥0.5, and: the absolute value of the maximum or minimum value of the change in the nut rotation angle is greater than 1°, and the growth coefficient P(d) of the average dynamic strain is greater than 0.5; a bolt loosening early warning is issued; the coefficient P(d) is used to eliminate false early warnings caused by insufficient initial samples, and to exclude interference data when the change in the nut rotation angle is relatively small. Generally, if the absolute value of the nut rotation angle is less than 1°, it can be considered that there is no risk of loosening.

[0117] like Figure 3 , 4 A bolt loosening monitoring device includes a bolt dynamic strain sensing module, a nut rotation angle sensing module, and a data acquisition and analysis module. The data acquisition and analysis module is connected to the bolt dynamic strain sensing module and the nut rotation angle sensing module.

[0118] The bolt dynamic strain sensing module is used to monitor the dynamic strain value of the flange bolts; the bolt dynamic strain sensing module is a dynamic strain gauge 2;

[0119] The nut rotation angle sensing module is used to monitor the rotation angle value of the nut relative to the bolt; the nut rotation angle sensing module is angle sensor 3;

[0120] The data acquisition and analysis module is used to acquire signals from the bolt dynamic strain sensing module and the nut rotation angle sensing module.

[0121] Specifically, the bolt dynamic strain sensing module is a full-bridge resistive dynamic strain gauge, which is welded to the bolt 1 at both ends; the nut rotation angle sensing module is fastened to the nut 4.

[0122] Specifically, the data acquisition and analysis module extracts the feature values ​​of each monitoring parameter online, performs coupling operations on the acquired signals and calculates the mean to obtain the mean sequence, and obtains T, S, and K; the bolt loosening probability assessment and early warning module is used to calculate P(C).

[0123] Specifically, the bolt loosening probability assessment and early warning module is used to assess the probability and trend of bolts loosening, and to issue loosening early warning information for bolts with a high probability of loosening.

[0124] Specifically, the bolt loosening degree and safety alarm module is used to assess the impact of bolt loosening and issue alarm information on the danger of loosening.

[0125] Specifically, such as Figure 2 t / n is 1 second, t is one minute, and n is 60; there are K mt time intervals within one month;

[0126] By continuously acquiring and coupling signals, the probability of an increasing trend in the mean dynamic strain sequence was obtained as: P(A) = S / K; the probability of an increasing trend in the mean nut rotation angle sequence was: P(B) = T / K; the bolt loosening coefficient was P(C) = 0.8P(A) + 0.2P(B). Dynamic strain data is more sensitive to the early stages of bolt loosening, while angle data is less sensitive. Based on the sensitivity of the two monitoring parameters to bolt loosening, different weighting coefficients can be determined. In this case, the weighting coefficient for dynamic strain data is 0.75, and for angle data it is 0.25. For different bolts, the coefficients can be iteratively corrected based on the monitoring results, thus correcting the coefficients for the corresponding bolts.

[0127] like Figure 1 Module 1 is a bolt dynamic strain sensing module, used to sense the dynamic strain value of flange bolts in real time. This module can be composed of multiple dynamic strain sensors. Multiple bolts can be selected at equal intervals for dynamic strain monitoring on each flange, and one dynamic strain gauge can be vertically installed on each bolt.

[0128] The performance indicators of the dynamic strain sensor must meet the following requirements:

[0129] Effective frequency band better than 0-10Hz; sensitivity greater than 0.1 mV / microstrain; maximum measurement range better than 3000 microstrains; resolution better than 1 microstrain; RS485 communication; dynamic strain gauges can be selected from commercially available full-bridge resistance dynamic strain gauges, such as the QZDB-B type dynamic strain gauge; the two ends of the strain gauge base are directly welded to bolts, such as... Figure 3 As shown.

[0130] Module 2 is a nut rotation angle sensing module, used to sense the rotation angle value of the nut relative to the bolt in real time; this module can be composed of multiple angle sensors. Each flange can select multiple bolts at equal intervals as the objects for nut loosening angle monitoring, and one angle sensor can be installed on each bolt;

[0131] The angle sensor's performance specifications must meet the following requirements: angle resolution better than 0.1 degrees; maximum measurement range of 360 degrees; RS485 communication; the inner dimensions of the magnetic base at the bottom of the angle sensor must match the outer dimensions of the bolt for screw fixing; the sensor's outer cover must be fixed to the upper surface of the nut, such as... Figure 4 As shown.

[0132] Module 3 is a data acquisition and analysis module, used to acquire dynamic strain signals, nut rotation angle signals, and extract the characteristic values ​​of each monitoring parameter online; the coupling operation used is used to screen valid data and eliminate interference; the correction coefficient β established for the bolt is used to calculate P(C), P(A), P(B); and P(d) is calculated.

[0133] Module 4 is the bolt loosening probability assessment and early warning module, which is used to assess the probability and trend of bolts loosening, and to issue loosening early warning information for bolts with a high probability of loosening;

[0134] Module 5 is a bolt loosening degree and safety alarm module, used to assess the impact of bolt loosening and issue a loosening hazard alarm.

[0135] Specifically, this invention also proposes a method for estimating the risk of loosening of multiple bolts.

[0136] Step 1: The degree of looseness of a single bolt is divided into: normal, looseness warning, and looseness alarm.

[0137] P(C) < 0.5, bolt looseness is normal;

[0138] P(C)≥0.5, loosening warning.

[0139] When conditions 1 and 2 are met, a bolt loosening alarm is triggered:

[0140] Condition 1: The absolute value of the change in the nut's rotation angle is greater than 1°;

[0141] Condition 2: The growth coefficient P(d) of the mean dynamic strain is greater than 0.5.

[0142] Step 2: The safety level for bolt loosening of the entire flange is determined according to the table below;

[0143]

[0144] In summary, this application is used and implemented as follows: A bolt monitoring signal acquisition and processing method is employed, using coupling operations to eliminate interference signals and obtain a reliable signal mean. Combined with this method, a bolt loosening coefficient calculation method is obtained, calculating the iterative correction coefficient β to determine the bolt loosening coefficient P(C), thereby obtaining a bolt pre-loosening early warning method. Furthermore, a method for calculating the growth coefficient P(d) of the bolt dynamic strain mean is proposed and used as an effective indicator for early warning activation. Applying the above data processing and jointly acquiring the required signals, a bolt loosening monitoring device is obtained, addressing the shortcomings of current wind turbine flange bolt loosening monitoring methods and devices.

[0145] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention.

Claims

1. A method for early warning of bolt loosening, characterized in that, Includes the following steps: Bolt monitoring signal acquisition and processing: After coupling the nut rotation angle and dynamic strain change value acquired within time period t, extract the corresponding acquired value whose coupling result is greater than or equal to 0, and remove the acquired value whose coupling result is less than 0; the coupling operation is: multiply the nut rotation angle and dynamic strain change value acquired within time period t one by one; At least one set of A and B was collected within the time period t; A represents a sequence of n numerical changes in the dynamic strain of the bolt relative to its initial value: ; B represents a sequence of n numerical values ​​showing the change in the nut's rotation angle relative to its initial value. ; Sequence A multiplied by sequence B equals sequence C: ; n≥z≥1, and z is an integer, satisfying the condition in the C sequence. Values ​​greater than or equal to 0: extract To obtain the data sequence ; extract To obtain the data sequence ; For data sequences , The mean values ​​are calculated separately to obtain the mean dynamic strain of the bolt and the mean rotation angle of the nut during time interval t, which are the characteristic values ​​of that time interval t. , ; From the first t-period to the m-th t-period, a sequence containing m feature values ​​was obtained through continuous monitoring: Dynamic strain mean sequence, ; The mean sequence of nut rotation angles ; Bolt loosening coefficient calculation: Sample sequence obtained within K time intervals (mt): ， ； A continuously increasing sequence: each value in the sequence is not less than its next adjacent value. That is, each value in the sequence satisfies: ; in, The number of consecutively increasing sequences in the sequence is denoted as S; in, The number of consecutively increasing sequences in the sequence is denoted as T; Probability calculations for assessing growth trends are performed on the eigenvalue sequences respectively: The probability that the dynamic strain mean sequence shows an increasing trend is: P(A) = S / K; The probability that the mean sequence of nut rotation angles shows an increasing trend is: P(B) = T / K; Bolt loosening coefficient P(C) = βP(A) + (1-β)P(B), 0≤β≤1, where β is a correction coefficient; Substituting P(A) and P(B), P(C) = (βS + T - βT) / K; Bolt loosening warning: If P(C) < 0.5, the bolt is safe and no warning is issued; if P(C) ≥ 0.5, a bolt loosening warning is issued.

2. The bolt pre-loosening early warning method according to claim 1, characterized in that, This includes calculating the growth factor of the mean dynamic strain of bolts: For a dynamic strain mean sequence, or more dynamic strain mean sequences connected in time sequence, a mean growth screening is performed. Let the sample of the obtained dynamic strain mean sequence be... ; sequence In any two adjacent values, i.e., in a sequence where h increases sequentially from 1 to p-1, the condition does not meet the requirement. Required, remove Filtering sequences The sequence was then obtained. ; For sequence Repeatedly perform mean-growth filtering until the sequence... All means in the range conform to ; sequence If the number of mean values ​​in the sample is r, then the growth coefficient is: P(d) = r / p; The growth coefficient of the average dynamic strain of the bolt is used to calculate P(d) and P(C) ≥ 0.

5. Furthermore, if the absolute value of the maximum or minimum value of the change in the nut rotation angle is greater than 1° and the growth coefficient of the average dynamic strain P(d) is greater than 0.5, a bolt loosening warning is issued.

3. A bolt loosening monitoring device, employing the bolt pre-loosening early warning method as described in claim 1 or 2, characterized in that, It includes a bolt dynamic strain sensing module, a nut rotation angle sensing module, and a data acquisition and analysis module. The data acquisition and analysis module is connected to the bolt dynamic strain sensing module and the nut rotation angle sensing module. The bolt dynamic strain sensing module is used to monitor the dynamic strain value of flange bolts; The nut rotation angle sensing module is used to monitor the rotation angle value of the nut relative to the bolt; The data acquisition and analysis module is used to acquire signals from the bolt dynamic strain sensing module and the nut rotation angle sensing module.

4. The bolt loosening monitoring device according to claim 3, characterized in that, The bolt dynamic strain sensing module is a full-bridge resistive dynamic strain gauge, which is welded to the bolt at both ends; the nut rotation angle sensing module is fastened to the nut.

5. A bolt loosening monitoring device according to claim 3, characterized in that, The data acquisition and analysis module extracts the feature values ​​of each monitoring parameter online. The data acquisition and analysis module performs coupling operation on the acquired signal and calculates the mean value to obtain the mean value sequence, and obtains T, S, and K as described in claim 1. A bolt loosening probability assessment and early warning module is used to calculate and obtain P(C) as described in claim 1.

Images