Method for judging tightening quality of GIS flange bolt based on torque-angle curve analysis

Abstract

The present application belongs to the technical field of data processing, and particularly relates to a GIS flange bolt tightening quality discrimination method based on torque rotation curve analysis, which comprises the following steps: resampling torque and rotation angle data based on a preset angle step, and calculating instantaneous tangent stiffness by using a Gaussian kernel weighted difference operator to smooth and suppress high-frequency friction noise in the aluminum alloy tightening process; constructing a dynamic elastic reference value based on linear elastic zone characteristics and hardening coefficient; calculating damage cumulative value based on the deviation of instantaneous stiffness and reference value, and determining micro yield and stopping when the cumulative value exceeds the threshold value. The present application effectively distinguishes transient viscous sliding interference and real plastic deformation through the energy integration idea, solves the misjudgment problem caused by the absence of obvious yield point and individual differences of cast aluminum alloy, and realizes high-precision nondestructive discrimination of bolt tightening quality.

Description

Technical Field

[0001] This invention relates to the field of data processing technology. More specifically, this invention relates to a method for judging the tightening quality of GIS flange bolts based on torque-angle curve analysis. Background Technology

[0002] Gas-insulated metal-enclosed switchgear (GIS) is a critical transmission and control device in power systems. Its housing is typically made of cast aluminum alloy and filled with high-pressure insulating gas. Therefore, the tightening quality of the flange bolts directly determines the sealing performance and operational safety of the equipment. Because cast aluminum alloy lacks a distinct yield point, and due to the presence of trace amounts of cutting fluid or metal debris during assembly, high-frequency viscous sliding friction easily occurs between the threads. This presents a technical challenge for accurately judging the tightening quality.

[0003] Existing tightening quality judgment technology usually adopts the torque-angle method. The main method is to calculate the instantaneous slope of the torque-angle curve, i.e., stiffness, based on the data of adjacent sampling points, and compare it with a preset fixed threshold to determine whether it is qualified.

[0004] However, this method of calculating stiffness based on two-point difference and relying on a fixed threshold has significant drawbacks in the GIS flange tightening scenario: On the one hand, the two-point difference algorithm is extremely sensitive to noise, which amplifies the instantaneous frictional force jump caused by viscous sliding through differential operation, forming a violently jagged waveform, making it difficult for the algorithm to distinguish between frictional interference and actual stiffness reduction, thus causing false alarms; on the other hand, the slope decay of aluminum alloys when micro-yielding is extremely weak and is often masked by the natural hardening trend of the material. Fixed parameter standards cannot identify this early plastic deformation, nor can they adapt to individual fluctuations caused by differences in shell material hardness and processing texture, which can easily lead to stripped bolts being misjudged as qualified. Summary of the Invention

[0005] To address the technical problems of lacking a clear yield point and being susceptible to viscous friction interference in the tightening of GIS cast aluminum alloy flange bolts, this invention provides a method for judging the tightening quality of GIS flange bolts based on torque-angle curve analysis. The method includes: acquiring torque and angle data during the bolt tightening process; resampling the angle data based on a preset angle step size to obtain an angle domain sequence, and acquiring the torque value corresponding to each sampling point in the angle domain sequence; and using a Gaussian kernel weighted difference operator to analyze the angle domain sequence and the corresponding torque values ​​within a sliding window containing multiple sampling points. The system performs calculations to determine the instantaneous tangential stiffness at each sampling point; determines the linear elastic zone based on the torque value; constructs a dynamic elastic reference value for sampling points beyond the linear elastic zone based on the arithmetic mean of all instantaneous tangential stiffnesses within the linear elastic zone and the rate of change of the instantaneous tangential stiffness; calculates a cumulative damage value based on the instantaneous tangential stiffness and the dynamic elastic reference value, wherein the increment of the cumulative damage value is positively correlated with the attenuation of the torque value and the instantaneous tangential stiffness; and determines that micro-yield has occurred and triggers a shutdown when the cumulative damage value exceeds a preset cumulative damage threshold.

[0006] This invention unifies the data dimension through angle domain resampling, eliminating non-physical noise introduced by motor speed fluctuations; it uses a Gaussian kernel weighted differential operator to calculate stiffness, which, compared to the traditional differential method, can effectively suppress viscous sliding friction noise during aluminum alloy tightening and avoid signal jaggedness; it constructs a dynamic elastic reference value, which can adapt to the natural hardening trend during flange flattening and avoid misjudgment caused by nonlinear stiffness growth; based on the discrimination mechanism of damage accumulation value, it uses the integral effect to distinguish between transient friction interference and continuous micro-yield, solving the detection problem caused by the lack of obvious yield point and susceptibility to friction interference in cast aluminum alloy materials, and significantly reducing the false alarm rate while ensuring high sensitivity.

[0007] Preferably, the formula for calculating the instantaneous tangential stiffness at the sampling point is: In the formula, For the first Instantaneous tangential stiffness at each sampling point; Let be the radius of the sliding window, and ; For data points within the sliding window relative to the first... The index offset of each sampling point; For the first Torque values ​​at each sampling point; For the first The arithmetic mean of the torque values ​​of all sampling points within the sliding window corresponding to each sampling point; For the first The angle value of each sampling point; For the first The arithmetic mean of the angle values ​​of all sampling points within the sliding window corresponding to each sampling point; For the first Gaussian weighting coefficients for each sampling point.

[0008] This invention utilizes the attenuation characteristics of Gaussian distribution to give higher weight to data near the center point of the sliding window. While preserving the trend of stiffness change, it effectively suppresses random friction noise far from the center, thereby outputting a smooth stiffness spectrum sequence with minimal phase lag, which significantly improves the signal-to-noise ratio and accuracy of stiffness calculation.

[0009] Preferably, the Gaussian weight coefficient of the sampling point is obtained by using the index offset of each data point within the sliding window relative to the center point as the independent variable.

[0010] Preferably, determining the linear elastic zone based on the torque value includes: scanning the angle domain sequence and selecting the range where the torque value is between 30% and 60% of the target value as the linear elastic zone.

[0011] This invention defines the cutoff range of the linear elastic zone. By selecting the range of 30% to 60% of the target torque value as the calculation basis, it can effectively avoid the unstable thread alignment stage in the early stage of tightening and the nonlinear hardening stage in the later stage. This ensures that the calculated average stiffness and reference parameters can accurately reflect the basic mechanical properties of the bolted connection and provide a reliable benchmark for subsequent judgment.

[0012] Preferably, the formula for calculating the dynamic elastic reference value of the sampling point is: In the formula, For the first Dynamic elastic reference value at each sampling point , This is the sequence number of the last sampling point in the linear elastic region; It is the arithmetic mean of the instantaneous tangent stiffness at all sampling points within the linear elastic region; For the first Instantaneous tangential stiffness at each sampling point; For the first The angle value of each sampling point; This is the ending angle value of the linear elastic zone; This is the hardening coefficient.

[0013] When constructing the dynamic elastic reference value, this invention introduces a hardening coefficient and a gain term for stiffness variation with angle, simulating the physical law of natural stiffness growth of GIS flange face during compaction, generating a reference curve that rises slightly with increasing angle, thereby dynamically adapting to the differences in material hardness and processing texture of different samples, effectively eliminating false stiffness attenuation false alarms caused by ignoring the natural hardening characteristics of materials.

[0014] Preferably, the cumulative damage value is equal to the sum of the cumulative damage value at the previous sampling point and the increment of the cumulative damage value; the increment of the cumulative damage value is equal to the product of the activation factor, torque value, and radian value converted from the preset standard angle step size at the sampling point, wherein the standard angle step size ranges from 0.1 degrees to 1.0 degrees.

[0015] This invention uses an iterative accumulation method to calculate the cumulative damage value, taking the magnitude of torque and the change in angle as the weighting terms of the increment. This means that the stiffness decay under high torque conditions represents a greater energy release and potential risk of damage, thereby more accurately quantifying the degree of plastic deformation of the bolted connection structure from an energy perspective and improving the physical meaning and reliability of the discrimination index.

[0016] Preferably, the method for obtaining the activation factor at the sampling point is as follows: calculate the relative stiffness attenuation rate at each sampling point based on the dynamic elastic reference value and the instantaneous tangential stiffness; compare the relative stiffness attenuation rate with a preset noise threshold: when the relative stiffness attenuation rate is less than the noise threshold, set the activation factor to zero; otherwise, the activation factor is exponentially positively correlated with the relative stiffness attenuation rate; the noise threshold ranges from 0.02 to 0.05.

[0017] This invention introduces an activation factor mechanism based on a noise threshold. By setting tiny stiffness fluctuations to zero, it effectively filters out background mechanical noise and non-destructive elastic fluctuations. Signals exceeding the threshold are amplified exponentially, enabling the algorithm to maintain extremely high sensitivity to real yield damage and achieving the best balance between anti-interference capability and detection sensitivity.

[0018] Preferably, the method for obtaining the relative stiffness attenuation rate at the sampling point is as follows: calculate the difference between the dynamic elastic reference value and the instantaneous tangential stiffness at the sampling point, then the relative stiffness at the sampling point is equal to the ratio of the difference to the dynamic elastic reference value at the sampling point.

[0019] Preferably, the target value of the torque and the hardening coefficient are obtained by performing tightening tests on samples from the same batch. Specifically, the method is as follows: the moment when the slope of the torque-angle curve of the sample first shows a significant decrease is obtained, and the torque value corresponding to this moment is recorded as the yield limit torque of the sample; the average yield limit torque of all samples is calculated, and the target value of the torque is set as a certain proportion of the average yield limit torque, with the proportion being 60% to 75%; the trend line of the instantaneous tangential stiffness of the sample during the normal compaction stage is obtained as a function of the rotation angle; the ratio of the slope of the trend line to the average instantaneous tangential stiffness of the sample is determined as the hardening coefficient.

[0020] Preferably, the method for determining the damage accumulation threshold is as follows: perform a tightening test on samples in the same batch until yielding occurs; calculate the damage accumulation value of each sample at the yield point; select the minimum value among the critical damage accumulation values ​​of all samples, and set the damage accumulation threshold as a certain proportion of the minimum value, which is 70% to 85%.

[0021] The beneficial effects of this invention are as follows:

[0022] This invention utilizes a Gaussian kernel-weighted difference operator to simultaneously suppress high-frequency friction noise during stiffness calculation; it constructs a dynamic elastic reference value that varies with angle to adapt to the natural hardening trend of aluminum alloy materials during compaction; and it introduces a damage energy integration mechanism to effectively distinguish between transient friction interference and true microscopic yielding through the cumulative stiffness attenuation effect. This method effectively reduces false alarms while maintaining high sensitivity, achieving early non-destructive quality assessment. Attached Figure Description

[0023] Figure 1 This is a flowchart illustrating the GIS flange bolt tightening quality judgment method based on torque-angle curve analysis in this invention;

[0024] Figure 2 This is a schematic diagram showing the comparison of stiffness calculation results;

[0025] Figure 3 This is a schematic diagram illustrating the instantaneous tangential stiffness and dynamic elastic reference values;

[0026] Figure 4 This is a schematic diagram illustrating the damage accumulation discrimination. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0029] This invention discloses a method for judging the tightening quality of GIS flange bolts based on torque-angle curve analysis, referring to... Figure 1 This includes steps S1 to S4:

[0030] S1: Acquire torque and angle data during bolt tightening; resample the angle data based on a preset angle step size to obtain an angle domain sequence, and obtain the torque value corresponding to each sampling point in the angle domain sequence.

[0031] It should be noted that the servo motor of the intelligent tightening gun does not rotate at a constant speed during operation, and usually shows a trend of being fast at first and then slowing down. This results in the data points collected based on the time dimension being sparse in the high-speed segment and dense in the low-speed segment. If subsequent differential calculations are performed directly based on the time series, non-physical numerical noise will be introduced due to the inconsistency of the sampling interval, thereby masking the weak material deformation characteristics. Therefore, this invention adopts the angle domain reconstruction method to unify the data dimension.

[0032] Specifically, the system first uses the high-frequency torque sensor and angle encoder built into the intelligent tightening gun to sample at a frequency... Real-time acquisition of raw data streams, including timestamps. Instantaneous torque value and instantaneous rotation angle value.

[0033] In this embodiment, the sampling frequency The sampling frequency is set to 2000 Hz; in other embodiments, the implementer can adjust the sampling frequency according to the controller's processing capabilities. Set to any value between 1000 Hz and 5000 Hz.

[0034] Furthermore, a standard angle step size is set. The step size is 0.5 degrees. This step size is chosen based on the accuracy considerations of GIS flange threads, ensuring that subtle changes in the thread are captured without causing data redundancy. In other embodiments, implementers can adjust the standard angle step size according to the precision requirements of the thread specifications. Set to 0.1 degrees to 1.0 degrees.

[0035] Finally, based on the standard angle step size This paper describes a method for reconstructing a non-uniformly spaced original data stream using cubic spline interpolation. The algorithm constructs a cubic polynomial between each adjacent data point, ensuring the continuity of the first and second derivatives of the interpolation function at the nodes. This results in a smooth, equally spaced angle domain sequence containing a standardized rotation sequence. and the corresponding torque sequence This eliminates the influence of motor speed fluctuations and purely reflects mechanical characteristics.

[0036] S2: Using the Gaussian kernel weighted difference operator, the angle domain sequence and the corresponding torque value are calculated within a sliding window containing multiple sampling points to determine the instantaneous tangential stiffness at each sampling point.

[0037] It should be noted that irregular micro-chips are generated during the tightening process of aluminum alloy materials, which causes high-frequency oscillations in the friction coefficient between threads. If the traditional two-point difference method is used to calculate stiffness, this friction noise will be amplified by the differential operation, forming a violent sawtooth waveform, making it impossible for the algorithm to distinguish between friction interference and actual stiffness reduction. Therefore, this invention introduces a Gaussian kernel weighted differential operator to smooth and suppress noise while calculating stiffness.

[0038] Specifically, based on the angle domain sequence, the instantaneous tangent stiffness at each sampling point is constructed, and the specific calculation formula is as follows:

[0039]

[0040] In the formula, For the first The instantaneous tangential stiffness at each sampling point is expressed in N·m / ° (Newton-meter per degree). The radius of the sliding window is 5, representing the range of data points involved in the calculation; in other embodiments, the implementer can adjust the sliding window radius according to the noise intensity. Set to an integer between 3 and 8; The data points within the sliding window are relative to the center point, i.e., the... The index offset of each sampling point; For the first The torque value at the sampling point, i.e., the first sampling point. Within the sliding window corresponding to the sampling point, the first... The torque value at each offset position; For the first The arithmetic mean of the torque values ​​of all sampling points within the sliding window corresponding to each sampling point; For the first The corner value of the nth sampling point, i.e., the nth Within the sliding window corresponding to the sampling point, the first... The angle value at each offset position; For the first The arithmetic mean of the angle values ​​of all sampling points within the sliding window corresponding to each sampling point; For the first Gaussian weighting coefficients for each sampling point.

[0041] Among them, Gaussian weight coefficients The formula for calculation is:

[0042]

[0043] In the formula, The data points within the sliding window are relative to the center point. The index offset; It is a natural exponential function; is the standard deviation of the Gaussian distribution, with a value of 2.0.

[0044] Among them, with Compared to As the magnitude of change increases, the value of the numerator increases, and the calculated... The increase reflects the faster rate of torque increase with steering angle; this is achieved by introducing a Gaussian weighting coefficient. Distance from the center point The closer the data point, the greater its weight; the farther away the data point, the smaller its weight. This calculation model can suppress random friction noise far from the center by utilizing the attenuation characteristics of Gaussian distribution while preserving the trend of stiffness change, and output a smooth stiffness spectrum sequence with minimal phase lag.

[0045] For example, the comparison chart of stiffness calculation results is as follows: Figure 2 As shown; among them, the stiffness calculated by the traditional two-point difference method is extremely sensitive to viscous sliding friction noise, and its waveform exhibits violent sawtooth oscillations, making it impossible to identify the true stiffness changes; for the stiffness calculated by the Gaussian kernel weighted difference operator in this invention, since the noise is effectively suppressed by using Gaussian weights while calculating the stiffness, a smooth and clear stiffness curve reflecting the tightening process is obtained, laying the foundation for subsequent calculations.

[0046] S3: Determine the linear elastic zone based on the torque value; construct the dynamic elastic reference value of the sampling points beyond the linear elastic zone based on the arithmetic mean of all instantaneous tangential stiffness within the linear elastic zone and the rate of change of instantaneous tangential stiffness.

[0047] It should be noted that, due to the micro-roughness of the GIS flange surface, as the bolt tightening force increases, the contact surface is gradually flattened, resulting in a natural increase in the actual stiffness of the connection structure and a nonlinear hardening trend. If a fixed stiffness threshold is used as a benchmark, the normal hardening phenomenon in the later stage of tightening will be misjudged as abnormal or the slight yield attenuation will be masked. Therefore, this invention constructs a dynamically increasing benchmark model through self-learning parameters.

[0048] First, the target value of the torque and the hardening coefficient are calibrated through experiments. The specific process is as follows:

[0049] 1. Select several aluminum alloy shell screw hole samples and matching bolts that are consistent with the actual production material and batch; use a high-frequency intelligent tightening gun to perform destructive tightening tests on all samples, that is, continuously tighten the bolts at the actual production speed until the bolt breaks or the internal thread completely slips out, causing the torque to be unloaded; during the test, record the torque, rotation angle and time data of each sample to form a test dataset.

[0050] 2. Determine the physical limits and target values ​​of torque, including: analyzing each curve in the test dataset, identifying the yield point, which is defined as the moment when the slope of the torque angle curve first shows a significant decrease, recording the torque value corresponding to this moment as the yield limit torque, and calculating the average yield limit torque of all samples; setting the target value of torque as a certain proportion of this average yield limit torque, usually taking a proportion of 60% to 75%, to ensure that the sealing requirements are met while reserving sufficient safety margin to prevent the material from approaching its failure limit during normal tightening.

[0051] 3. Extract normal characteristics to determine the hardening coefficient, including: extracting the healthy state segment representing the bolt under normal compaction and without yielding from each test curve; performing linear trend analysis on the instantaneous tangential stiffness data within the healthy state segment to obtain the trend line of instantaneous tangential stiffness changing with the increase of rotation angle; determining the ratio of the growth rate of the trend line to the average instantaneous tangential stiffness of the healthy state segment as the hardening coefficient, which reflects the physical characteristics of the natural stiffness increase of the current batch of aluminum alloy material during the compaction process.

[0052] Then, the angle domain sequence is scanned, and the range of torque values ​​between 30% and 60% of the target value is selected as the linear elastic region; the arithmetic mean of the instantaneous tangential stiffness at all sampling points within the linear elastic region is calculated.

[0053] Furthermore, a dynamic elastic reference value for the current bolt after it exceeds the linear elastic region is constructed. The specific calculation formula is as follows:

[0054]

[0055] In the formula, For the first Dynamic elastic reference value at each sampling point , This is the sequence number of the last sampling point in the linear elastic region; It is the arithmetic mean of the instantaneous tangential stiffness at all sampling points within the linear elastic region, representing the basic hardness of the bolt; For the first Instantaneous tangential stiffness at each sampling point; For the first The angle value of each sampling point; This is the ending angle value of the linear elastic zone; This is the hardening coefficient.

[0056] in, As a gain term, the value of the gain term increases linearly with the increase of the rotation angle, causing the value of the dynamic elastic reference value to fluctuate within the base hardness range. The hardening coefficient is gradually increased from the base level; this calculation model uses the hardening coefficient... The physical hardening process of the flange surface during compaction was simulated, and a reference curve that rises slightly with the increase of angle was generated, thus avoiding misjudgment caused by ignoring the natural hardening characteristics of the material.

[0057] For example, a schematic diagram of the instantaneous tangential stiffness and dynamic elastic reference values ​​is shown below. Figure 3 As shown, the curve represents the linear elastic zone where the torque reaches 30%-60% of the target value, automatically identified by the system. The curve for the dynamic elastic reference value, constructed based on the average stiffness and preset hardening coefficient of this zone, shows a slight upward movement with increasing angle, reflecting the normal physical compaction and hardening process. Regarding the curve corresponding to the instantaneous tangential stiffness, in the later stages of tightening, the instantaneous tangential stiffness begins to fall significantly below the dynamic elastic reference value, which is a signal of micro-yield. Therefore, by calculating the instantaneous tangential stiffness and the dynamic elastic reference value, the problem of the traditional fixed threshold method being unable to adapt to the natural hardening of materials is solved.

[0058] S4: Calculate the damage accumulation value based on the instantaneous tangential stiffness and dynamic elastic reference value. The increment of the damage accumulation value is positively correlated with the torque value and the degree of attenuation of the instantaneous tangential stiffness. In response to the damage accumulation value exceeding the preset damage accumulation threshold, it is determined that micro-yield has occurred and the machine is triggered to stop.

[0059] It should be noted that the stiffness decrease caused by friction interference is usually transient and will recover rapidly, while the stiffness decrease caused by micro-yielding is continuous and irreversible. It is difficult to distinguish between the two based on the stiffness difference at a certain moment, resulting in a very high false alarm rate when using high-sensitivity detection. Therefore, this invention introduces the concept of energy integration to calculate the cumulative effect of stiffness deviation.

[0060] Specifically, based on the dynamic elastic reference value and the instantaneous tangential stiffness, the relative stiffness attenuation rate at each sampling point is calculated, and the specific calculation formula is as follows:

[0061]

[0062] In the formula, For the first The relative stiffness attenuation rate at each sampling point; For the first Dynamic elastic reference value at each sampling point; For the first Instantaneous tangential stiffness at each sampling point.

[0063] Furthermore, based on the relative stiffness attenuation rate, the activation factor at each sampling point is calculated, and the specific calculation formula is as follows:

[0064]

[0065] In the formula, For the first The activation factor at each sampling point is used to amplify significant damage signals; For the first The relative stiffness attenuation rate at each sampling point; The noise threshold is set to 0.03. In other embodiments, the implementer can adjust the noise threshold according to the on-site vibration conditions. Set it to between 0.02 and 0.05; It is a natural exponential function.

[0066] Furthermore, the cumulative damage value at each sampling point is calculated using the following formula:

[0067]

[0068] In the formula, For the first The cumulative damage value at each sampling point, in joules; For the first The cumulative damage value at each sampling point; For the first Activation factors at each sampling point; For the first The torque value at each sampling point is used as a weighting term to indicate that stiffness decay under high torque means greater energy release; Standard angle step size The converted radian value, and .

[0069] Among them, when Less than hour, The cumulative damage value is 0. The fact that it remains unchanged indicates that small stiffness fluctuations are ignored; when Greater than or equal to At that time, with A tiny increase, The energy increment increases exponentially, and combined with the high weighting of the high torque value, this significantly increases the energy increment per step, leading to a surge in cumulative damage. Rapidly increasing; the calculation model can effectively filter transient friction noise through integral accumulation, responding only to continuous stiffness decay.

[0070] Furthermore, the test dataset is substituted into the method of the present invention for simulation operation, and the damage accumulation value of each test curve when it reaches the yield point is recorded as the critical damage accumulation value of each sample. The minimum value among the critical damage accumulation values ​​of all samples is selected, and the damage accumulation threshold is set as a certain proportion of the minimum value, usually 70% to 85%. This ensures that when the damage accumulation value in production reaches the threshold and triggers a shutdown, the state of the thread has not yet reached the real physical failure point, and only a small amount of repairable plastic deformation has occurred, thereby achieving non-destructive protection of the workpiece.

[0071] Ultimately, when the cumulative damage value When the set damage accumulation threshold is exceeded, micro-yield is determined to have occurred and a shutdown is triggered.

[0072] For example, a schematic diagram of damage accumulation discrimination is shown below. Figure 4 As shown in the figure; among them, the curve corresponding to the damage accumulation value is as follows: in the early stage of tightening, the damage value remains at 0 because the stiffness deviation is small and does not reach the noise threshold; when entering the micro-yielding stage, the actual stiffness is continuously lower than the benchmark, and the activation factor exponentially amplifies this deviation, causing the damage accumulation value to accumulate and rise rapidly; when the damage accumulation value exceeds the preset shutdown threshold at about 48.5 degrees, the system determines that micro-yielding has occurred and triggers the shutdown point; this discrimination method based on the cumulative effect is more robust and accurate than the threshold judgment at a single moment, and solves the problem of false alarms caused by transient friction interference.

Claims

1. A method for judging the tightening quality of GIS flange bolts based on torque-angle curve analysis, characterized in that, include: Acquire torque and angle data during bolt tightening; resample the angle data based on a preset angle step size to obtain an angle domain sequence, and obtain the torque value corresponding to each sampling point in the angle domain sequence; The Gaussian kernel weighted difference operator is used to calculate the angle domain sequence and the corresponding torque value within a sliding window containing multiple sampling points in order to determine the instantaneous tangential stiffness at each sampling point; The linear elastic zone is determined based on the torque value; the dynamic elastic reference value of the sampling points beyond the linear elastic zone is constructed based on the arithmetic mean of all instantaneous tangential stiffness within the linear elastic zone and the rate of change of instantaneous tangential stiffness. The cumulative damage value is calculated based on the instantaneous tangential stiffness and dynamic elastic reference value. The increment of the cumulative damage value is positively correlated with the torque value and the degree of attenuation of the instantaneous tangential stiffness. The cumulative damage value is equal to the sum of the cumulative damage value at the previous sampling point and the increment of the cumulative damage value; The increment of the damage accumulation value is equal to the product of the activation factor, torque value, and radian value converted from the preset standard angle step size at the sampling point. The standard angle step size ranges from 0.1 degrees to 1.0 degrees. The activation factor at the sampling point is obtained as follows: based on the dynamic elastic reference value and the instantaneous tangential stiffness, the relative stiffness at each sampling point is calculated; the relative stiffness attenuation rate is compared with the preset noise threshold; when the relative stiffness attenuation rate is less than the noise threshold, the activation factor is set to zero. Otherwise, the activation factor is exponentially positively correlated with the relative stiffness decay rate; the noise threshold ranges from 0.02 to 0.

05. The method for obtaining the relative stiffness attenuation rate at the sampling point is as follows: calculate the difference between the dynamic elastic reference value and the instantaneous tangential stiffness at the sampling point. The dynamic elastic reference value is related to the arithmetic mean of the instantaneous tangential stiffness. The relative stiffness attenuation rate at the sampling point is equal to the ratio of the difference to the dynamic elastic reference value at the sampling point. In response to the damage accumulation value exceeding the preset damage accumulation threshold, micro-yield is determined to have occurred and a shutdown is triggered.

2. The method for judging the tightening quality of GIS flange bolts based on torque-angle curve analysis according to claim 1, characterized in that, The formula for calculating the instantaneous tangential stiffness at the sampling point is: ； In the formula, For the first Instantaneous tangential stiffness at each sampling point; Let be the radius of the sliding window, and ; For data points within the sliding window relative to the first... The index offset of each sampling point; For the first Torque values ​​at each sampling point; For the first The arithmetic mean of the torque values ​​of all sampling points within the sliding window corresponding to each sampling point; For the first The corner value of each sampling point; For the first The arithmetic mean of the angle values ​​of all sampling points within the sliding window corresponding to each sampling point; For the first Gaussian weighting coefficients for each sampling point.

3. The method for judging the tightening quality of GIS flange bolts based on torque-angle curve analysis according to claim 2, characterized in that, The Gaussian weighting coefficient of the sampling point is obtained by using the index offset of each data point within the sliding window relative to the center point as the independent variable.

4. The method for judging the tightening quality of GIS flange bolts based on torque-angle curve analysis according to claim 1, characterized in that, The step of determining the linear elastic zone based on the torque value includes: scanning the angle domain sequence and selecting the range where the torque value is between 30% and 60% of the target value as the linear elastic zone.

5. The method for judging the tightening quality of GIS flange bolts based on torque-angle curve analysis according to claim 4, characterized in that, The formula for calculating the dynamic elastic reference value of the sampling point is: ； In the formula, For the first Dynamic elastic reference value at each sampling point , This is the sequence number of the last sampling point in the linear elastic region; It is the arithmetic mean of the instantaneous tangent stiffness at all sampling points within the linear elastic region; For the first Instantaneous tangential stiffness at each sampling point; For the first The corner value of each sampling point; This is the ending angle value of the linear elastic zone; This is the hardening coefficient.

6. The method for judging the tightening quality of GIS flange bolts based on torque-angle curve analysis according to claim 5, characterized in that, The target value of the torque and the hardening coefficient were obtained by performing tightening tests on samples from the same batch. The specific method is as follows: The moment when the slope of the torque-angle curve of the sample first shows a significant decrease is recorded as the moment when the torque value corresponding to that moment is the yield limit torque of the sample. The average yield limit torque of all samples is calculated, and the target value of the torque is set as a certain proportion of this average yield limit torque, with this proportion ranging from 60% to 75%. Obtain the trend line of the instantaneous tangential stiffness of the sample during the normal compaction stage as a function of rotation angle; determine the hardening coefficient as the ratio of the slope of the trend line to the average instantaneous tangential stiffness of the sample.

7. The method for judging the tightening quality of GIS flange bolts based on torque-angle curve analysis according to claim 1, characterized in that, The method for determining the damage accumulation threshold is as follows: Tightening tests were performed on samples from the same batch until yielding occurred; the cumulative damage value at the yield point of each sample was calculated; the minimum value among the critical cumulative damage values ​​of all samples was selected, and the cumulative damage threshold was set as a certain percentage of this minimum value, which was 70% to 85%.

Images