A method for acquiring and processing torque impact signals from PDC drill bits

Abstract

This invention belongs to the field of electrical digital data processing technology, specifically relating to a method for acquiring and processing torque impact signals from a PDC drill bit. The method includes: dividing torque data into several torque data windows; determining the torque waveform complexity of each torque data window; determining the rotational speed slip intensity and formation engagement complexity of each torque data window; determining the adaptive order of the AR model for each torque data window; performing predictive filtering on the torque data within each torque data window to obtain a prediction residual sequence for the torque data window; and extracting the torque impact signal based on the numerical characteristics of the prediction residual sequence. This invention overcomes the limitations of fixed-order AR models by analyzing the numerical characteristics of torque and rotational speed data and adaptively adjusting the order of the AR model, thereby filtering out background noise and improving the accuracy and robustness of the torque impact signal.

Description

Technical Field

[0001] This invention relates to the field of electrical digital data processing technology. More specifically, this invention relates to a method for acquiring and processing torque impact signals from a PDC drill bit. Background Technology

[0002] In the exploration and development of deep oil and gas resources, PDC drill bits have become a core tool in drilling engineering due to their efficient cutting and rock-breaking capabilities. As drilling depth increases, the downhole geological environment becomes extremely complex, with frequent alternations between soft and hard formations, and the slenderness ratio of the drill string increases significantly. This working environment causes PDC drill bits to easily exhibit complex dynamic behaviors at the bottom of the well, the most typical of which is stick-slip vibration. Stick-slip vibration is a nonlinear self-excited torsional vibration. The drill bit alternately experiences a viscous period and a slip period at the bottom of the well. During the viscous period, the drill bit speed is almost zero, and the torque accumulates rapidly due to the continuous rotation of the drill string. During the slip period, the accumulated elastic energy is released instantaneously, causing the drill bit speed to soar to several times the surface speed, accompanied by a violent release of torque. This high-energy, low-frequency, large-amplitude vibration background constitutes an extremely harsh noise environment.

[0003] However, the torsional impact signal generated by early chipping of the composite blades during the cutting process of PDC drill bits typically exhibits transient characteristics of weak energy, wide frequency band, and extremely short duration. These early fault signals, which are crucial for assessing the health of the drill bit, are easily drowned out by the strong stick-slip vibration background, making them unrecognizable by ground monitoring systems. Existing processing methods usually employ autoregressive (AR) models to linearly predict the torsional signal and extract the impact component by analyzing the prediction residuals. AR models can fit the background noise well, filtering it out from the original signal.

[0004] Existing AR models, when applied to downhole torque data processing, typically use a fixed model order. However, the downhole geological environment and drill string movement are dynamically coupled and constantly changing. AR models with a fixed order inherently lack the ability to adapt to such environmental changes. This fixed order setting cannot adaptively adjust the model's memory depth according to the actual complex formation conditions. It may not be suitable for homogeneous soft formations because such a pre-set high order can lead to overfitting of the model, resulting in the disappearance of impact features in the prediction residuals. For hard interlayers or heterogeneous formations, a low order can lead to underfitting of the model, failing to completely remove complex background noise and affecting the extraction accuracy of PDC torque impact signals. Summary of the Invention

[0005] To address the technical problem of traditional AR models failing to adapt to dynamic downhole geological environments when processing torque signals from PDC drill bits, leading to missed detections or misjudgments of weak impact signals due to their fixed order, this invention provides a method for acquiring and processing torque impact signals from PDC drill bits. The method includes: acquiring torque and rotational speed data from the PDC drill bit, and dividing the torque data into several torque data windows; determining the torque waveform complexity of the torque data window based on the mean of the numerical differences between adjacent data points within the torque data window and the standard deviation within the torque data window; determining the rotational speed slip intensity of the torque data window based on the rate of change of numerical values ​​between adjacent data points within the rotational speed data segment corresponding to the torque data window; determining the formation engagement complexity of the torque data window based on the torque waveform complexity and the rotational speed slip intensity; determining the adaptive order of the AR model for the torque data window based on the formation engagement complexity; performing predictive filtering on the torque data within the torque data window using the adaptive order to obtain a predictive residual sequence for the torque data window; and extracting the torque impact signal based on the numerical characteristics of the predictive residual sequence.

[0006] This invention identifies the preliminary complexity of formation lithology and mechanical interference by calculating the complexity of the torque waveform and utilizing time-domain fluctuation characteristics. By calculating the rotational slip intensity, it identifies the torque signal caused by mechanical factors. By comprehensively considering the complexity of the torque waveform and the rotational slip intensity, it can more accurately distinguish between the actual formation meshing complexity and spurious mechanical interference. By dynamically adjusting the adaptive order of the AR model according to the formation meshing complexity and performing predictive filtering, it achieves the matching of the model order to signals under different working conditions. This avoids underfitting of low-order models to complex formations and prevents overfitting of high-order models to homogeneous formations or slip stages. In complex downhole environments, it can extract torque impact signals more accurately, thereby improving the reliability of PDC drill bit health status assessment.

[0007] Preferably, the acquisition of torque and rotational speed data of the PDC drill bit includes: acquiring torque data of the downhole drill bit during rock cutting using a strain gauge torque sensor; and acquiring rotational speed data of the downhole drill bit during rock cutting using a MEMS gyroscope.

[0008] Preferably, dividing the torque data into several torque data windows includes: dividing the torque data into several torque data windows of equal length using a sliding window.

[0009] Preferably, the torque waveform complexity of the torque data window satisfies the expression:

[0010] In the formula, For the first The complexity of torque waveforms for each torque data window. The total number of data points within each torque data window. For the first The first torque data window Torque values ​​at each data point For the first The first torque data window Torque values ​​at each data point For the first The standard deviation of torque data within a single torque data window It is the maximum-minimum normalization function.

[0011] This invention constructs a composite function of the sum of the absolute values ​​of the differences between adjacent data points and the standard deviation, thereby enabling the assessment of torque waveform complexity. The adjacent difference term reflects the local roughness and high-frequency jumps of the signal, while the standard deviation term amplifies the influence of the overall fluctuation intensity. This allows the torque data window under complex formations or mechanical interference to calculate a larger torque waveform complexity, thus mapping the temporal texture characteristics of downhole signals and providing a reliable basis for calculating formation meshing complexity.

[0012] Preferably, the rotational slip strength of the torque data window satisfies the expression:

[0013] In the formula, For the first The speed-slip strength of each torque data window The total number of data points within each torque data window. For the first The torque data window corresponds to the first speed data segment within the same time period. The rotational speed values ​​at each data point For the first The torque data window corresponds to the first speed data segment within the same time period. The rotational speed values ​​at each data point For the first Each torque data window corresponds to the sampling time interval of the speed data segment within the same time period. It is the maximum-minimum normalization function.

[0014] This invention achieves the assessment of rotational slip strength by constructing a composite function that divides the sum of the absolute values ​​of the differences between adjacent rotational speed data points by the sampling time interval. The rotational speed difference term reflects the change in rotational speed, and the sampling time interval, as the denominator, together with the rotational speed difference term, reflects the rate of change of rotational speed. This allows the calculation of rotational speed data windows under mechanical slip to obtain a larger value of rotational slip strength, thereby mapping the drastic fluctuation characteristics of drill bit rotational speed and providing a reliable basis for calculating the complexity of formation meshing.

[0015] Preferably, the formation engagement complexity of the torque data window satisfies the expression:

[0016] In the formula, For the first Formation meshing complexity for each torque data window For the first The complexity of torque waveforms for each torque data window. For the first The speed-slip strength of each torque data window It is a natural exponential function.

[0017] This invention achieves the assessment of formation meshing complexity by constructing a composite function that includes torque waveform complexity and rotational speed slip strength index. The rotational speed slip strength index amplifies the weight of mechanical slip in interference determination through exponential decay mapping, so that the torque data window under mechanical slip can calculate a smaller formation meshing complexity, thereby enabling more effective identification and suppression of torque signals caused by mechanical factors.

[0018] Preferably, the adaptive order of the AR model of the torque data window satisfies the expression:

[0019] In the formula, For the first Adaptive order of the AR model for a torque data window This is the maximum order of the preset AR model. The minimum order of the preset AR model. For the first Formation meshing complexity for each torque data window This is the floor function.

[0020] This invention achieves adaptive determination of the order of the AR model by constructing a linear gain function based on the formation meshing complexity. The order range is dynamically adjusted by utilizing the formation meshing complexity. When complex formations are detected, the model order is stretched to enhance the background fitting ability, while when mechanical slippage or homogeneous formations are detected, the model order is shortened to retain impact details, ensuring that torsional impact signals can be effectively extracted while stripping away complex backgrounds.

[0021] Preferably, the step of predicting and filtering the torque data within the torque data window includes: estimating the autoregressive coefficients of the AR model of the torque data window based on the adaptive order of the AR model of the torque data window using the Burg algorithm, and performing linear prediction on the torque data within the torque data window using the AR model of the torque data window to obtain the predicted value of the torque data within the torque data window.

[0022] Preferably, the method for obtaining the prediction residual sequence of the torque data window is as follows: the difference between the torque data in the torque data window and the predicted value of the torque data in the torque data window is used as the prediction residual sequence of the torque data window.

[0023] Preferably, the extraction of the torsional impact signal includes: calculating the kurtosis value of the predicted residual sequence of the torsional data window; in response to the kurtosis value of the predicted residual sequence of the torsional data window being greater than a preset kurtosis threshold, a torsional impact signal exists in the torsional data window; and recording the time and amplitude of the torsional impact signal occurrence.

[0024] The beneficial effects of this invention are as follows:

[0025] This invention solves the technical problem of poor adaptability of traditional AR models with fixed orders in complex downhole environments by introducing an adaptive mechanism based on the formation meshing complexity of the torque data window to the order of the AR model.

[0026] This invention analyzes the temporal fluctuations of torque data and the rate of change of rotational speed data, and tightly couples the indicators reflecting the complexity of strata lithology with the indicators reflecting mechanical slippage, thereby realizing the identification and suppression of torque signals caused by mechanical factors.

[0027] This invention can calculate an adaptive order matching the operating characteristics of each torque data window. Through dynamic adjustment of the AR model, it achieves an accurate match between the order and signal complexity. For the mechanical slip stage, the order is automatically reduced to prevent overfitting; for complex formation stages, the order is increased to enhance background stripping, thereby improving the signal-to-noise ratio of weak torque impact signals in the predicted residual sequence. Ultimately, this invention improves the accuracy and robustness of torque impact signal extraction from PDC drill bits, enabling timely and accurate identification of early edge breakage faults, and providing a solid technical guarantee for the safety optimization and accident prevention of drilling projects. Attached Figure Description

[0028] Figure 1 This is a flowchart illustrating a method for acquiring and processing torque impact signals from a PDC drill bit according to the present invention.

[0029] Figure 2 The results of predictive filtering for AR models with a fixed lower order in existing technologies;

[0030] Figure 3 The results of predictive filtering for AR models with a fixed higher order in the existing technology;

[0031] Figure 4 The result of predictive filtering is shown for the adaptive order AR model of this invention. Detailed Implementation

[0032] 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.

[0033] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0034] This invention discloses a method for acquiring and processing torque impact signals from a PDC drill bit, referring to... Figure 1 This includes steps S001 to S005, specifically:

[0035] S001. Obtain the torque and rotation speed data of the PDC drill bit, and divide the torque data into several torque data windows.

[0036] Specifically, a strain gauge torque sensor and a MEMS gyroscope are installed near the drill bit sub of the PDC drill bit. Within the same time period, the strain gauge torque sensor acquires torque data of the downhole drill bit during rock-breaking, and the MEMS gyroscope acquires rotational speed data of the downhole drill bit during the same rock-breaking process. After acquiring the torque data, a sliding window is used to divide the torque data into several equal-length torque data windows. In this embodiment, the acquisition frequency of both torque and rotational speed data is 1000Hz. The torque data window is divided using a sliding window with a length of 1000 data points. In other embodiments, the implementer can set the acquisition frequency of torque and rotational speed data and the torque data window division method according to the actual implementation situation.

[0037] S002. Determine the torque waveform complexity of the torque data window based on the mean of the numerical differences between adjacent data points within the torque data window and the standard deviation within the torque data window.

[0038] It should be noted that when using traditional AR models to process PDC drill bit downhole torque data, the fixed order is difficult to adapt to changes in the downhole environment, which leads to a decrease in the extraction accuracy of PDC torque impact signals. According to the principle of drill bit rock breaking dynamics, the more complex the formation lithology, the more violent and disordered the torque data fluctuations during drill bit rock breaking, and the richer the high-frequency components in the signal. Therefore, this invention combines the numerical change characteristics between adjacent data points within the torque data window and the numerical discrete characteristics of the data to determine the torque waveform complexity of the torque data window, which is used to comprehensively characterize the degree of volatility of the torque signal under the current drilling state.

[0039] Specifically, the complexity of the torque waveform satisfies the expression:

[0040] ;

[0041] In the formula, For the first The complexity of torque waveforms for each torque data window. The total number of data points within each torque data window. For the first The first torque data window Torque values ​​at each data point For the first The first torque data window Torque values ​​at each data point For the first The standard deviation of torque data within a single torque data window It is the maximum-minimum normalization function.

[0042] In the formula, The larger the value, the more likely it is to be the first. The more drastic the jump in torque data values ​​within a given torque data window between adjacent time points, the more likely the torque data is to change. The richer the high-frequency texture of the signal within each torque data window, the better. The greater the complexity of the torque waveform in each torque data window. The larger the value, the more likely it is to be the first. The more pronounced the overall dispersion of the torque data values ​​within a given torque data window, the more significant the overall dispersion. The more drastic the jump in torque data values ​​within a given torque data window between adjacent time points, the greater the reliability, indicating that the [missing information - likely a specific type of torque data window]. The more complex the torque data fluctuations within each torque data window, the more complex the torque data fluctuations become. The greater the complexity of the torque waveform in each torque data window.

[0043] S003. Based on the rate of change of values ​​between adjacent data points in the speed data segment corresponding to the torque data window within the same time period, determine the speed slip strength of the torque data window, and determine the formation meshing complexity of the torque data window based on the torque waveform complexity and the speed slip strength.

[0044] It should be noted that during the slip phase of stick-slip vibration in a PDC drill bit, energy is released rapidly, causing the drill bit speed to spike instantaneously. This results in significant oscillations and jumps in the torque signal due to the violent movement of the mechanical system. Such mechanically driven fluctuations can be confused with the high-frequency textures generated by cutting hard formations. Matching a high-order AR model solely to the torque waveform fluctuations leads to overfitting the mechanical background waveform during the slip phase, thus filtering out weak, genuine torque impact signals and causing the loss of fault characteristics in the prediction residuals. However, mechanically driven fluctuations are usually accompanied by a spike in rotational speed, while the torque impact signal corresponds to a relatively stable rotational speed over the same time period. Therefore, this invention combines the numerical variation characteristics of rotational speed data with the complexity of the torque waveform to determine the formation engagement complexity of the torque data window. It uses the instantaneous rate of change of rotational speed data to characterize the mechanical slip intensity and corrects the torque waveform complexity to distinguish between genuine torque impact signals and spurious mechanical interference.

[0045] Specifically, the rotational slip strength satisfies the expression:

[0046] ;

[0047] In the formula, For the first The speed-slip strength of each torque data window This represents the total number of data points within each torque data window. Since the acquisition time period and frequency for torque and speed data are consistent across different time periods in this invention, the total number of data points within the speed data segment corresponding to each torque data window is also [the total number of data points]. , For the first The torque data window corresponds to the first speed data segment within the same time period. The rotational speed values ​​at each data point For the first The torque data window corresponds to the first speed data segment within the same time period. The rotational speed values ​​at each data point For the first Each torque data window corresponds to the sampling time interval of the speed data segment within the same time period. It is the maximum-minimum normalization function.

[0048] In the formula, It indicates the first Each torque data window corresponds to the instantaneous speed change rate within the same time period's speed data segment. The larger the value, the more likely it is to be the first. The more drastic the change in the torque data window within the same time period, the more it indicates that the torque data window corresponds to the speed data segment within the same time period. The more severe the mechanical slippage within the time period corresponding to each torque data window, the more severe the mechanical slippage becomes. The greater the speed slip strength in each torque data window, the higher the slip strength.

[0049] Furthermore, the formation meshing complexity satisfies the expression:

[0050] ;

[0051] In the formula, For the first Formation meshing complexity for each torque data window For the first The complexity of torque waveforms for each torque data window. For the first The speed-slip strength of each torque data window It is a natural exponential function.

[0052] In the formula, The larger the value, the more important it is to analyze the numerical characteristics of torque data. The more complex the torque data fluctuations within each torque data window, the more complex the torque data fluctuations become. The greater the formation meshing complexity within each torque data window, the better. The larger the value, the more likely it is to be the first. The more severe the mechanical slippage within the time period corresponding to each torque data window, the better for the first... The more pronounced the suppression of torque waveform complexity in the first torque data window, the better. The smaller the formation meshing complexity of each torque data window.

[0053] S004. Determine the adaptive order of the AR model for the torque data window based on the formation meshing complexity.

[0054] It should be noted that after obtaining the formation meshing complexity of the torque data window, this invention will calculate the adaptive order of the AR model of the torque data window based on the formation meshing complexity. Traditional AR models usually set a fixed order when processing torque signals. This invention, however, uses the formation meshing complexity to calculate the adaptive order of the AR model of the torque data window, so that the AR model can extract torque impact signals more accurately.

[0055] Specifically, the adaptive order satisfies the expression:

[0056] ;

[0057] In the formula, For the first Adaptive order of the AR model for a torque data window This is the maximum order of the preset AR model, for example. , For example, the minimum order of the preset AR model. , For the first Formation meshing complexity for each torque data window This is the floor function.

[0058] In the formula, The larger the value, the more likely it is to be the first. The more complex the formation where the PDC drill bit is located within the corresponding time period for each torque data window, the more complex the formation needs to be. To ensure that the AR model has sufficient ability to fit and filter out complex background noise, the first... The closer the adaptive order of the AR model for each torque data window is to the preset maximum order of the AR model, the better.

[0059] S005. The torque data in the torque data window is predicted and filtered using an adaptive order to obtain the prediction residual sequence of the torque data window, and the torque impact signal is extracted based on the numerical characteristics of the prediction residual sequence.

[0060] Specifically, extracting the torsional impact signal includes:

[0061] The Burg algorithm is used to estimate the autoregressive coefficients of the AR model based on the torque data window using the adaptive order of the torque data window. The torque data within the torque data window is then linearly predicted using the AR model with the adaptive order of the torque data window, thus obtaining the predicted value of the torque data within the torque data window.

[0062] Furthermore, the difference between the torque data in the torque data window and the predicted value of the torque data within the torque data window is calculated to obtain the prediction residual sequence of the torque data window;

[0063] Furthermore, the kurtosis value of the predicted residual sequence of the torque data window is calculated;

[0064] It should be noted that this invention uses kurtosis as a statistical indicator, which can keenly capture the transient impact characteristics hidden in the predicted residual sequence. Since the background residual after AR model filtering approximately follows a normal distribution, the corresponding kurtosis value is 3. The torsional impact signal caused by early blade breakage will increase the kurtosis value. This invention ensures that the torsional impact signal can be accurately detected by monitoring the sudden change in kurtosis value, avoiding missed detection due to background interference.

[0065] In response to the fact that the kurtosis value of the predicted residual sequence of the torque data window is greater than the preset kurtosis threshold, a torque impact signal exists in the torque data window. The time and amplitude of the torque impact signal are recorded, and then a blade breakage warning is issued to the ground technicians. In this embodiment, the kurtosis threshold is set to 3.5. In other embodiments, the implementers can set the kurtosis threshold according to the actual implementation situation.

[0066] like Figure 2 , Figure 3 and Figure 4 As shown, Figure 2 The results of prediction filtering using a fixed low-order AR model in existing technology are shown. In the time interval of 2s to 5s, the prediction residual sequence exhibits violent and chaotic fluctuations. This is because the fixed model order is set too low, resulting in insufficient memory depth of the model. It is unable to effectively capture and fit the rich high-frequency texture features under complex geological conditions, resulting in a serious underfitting phenomenon. A large amount of background noise is not filtered out and remains directly in the residual, forming a high-amplitude interference band. This background noise completely drowns out the torsional impact signal, making it impossible to distinguish the characteristic amplitude of the torsional impact signal from the background noise. This makes it impossible to identify the torsional impact signal and causes the early fault to be missed. Figure 3 The results of predictive filtering using an AR model with a fixed high order in existing technology are shown. Although the background noise throughout the time period is suppressed relatively flat and the filtering effect seems good, careful observation reveals that the torque impact signal that should have been present has almost completely disappeared. This is because the fixed model order is set too high, which leads to a serious model overfitting phenomenon. The overly strong fitting ability causes the algorithm to mistakenly filter out the torque impact signal as background noise, resulting in a significant attenuation of the torque impact signal amplitude, loss of key fault feature information, and loss of the ability to detect early faults. Figure 4 The results of predictive filtering using an adaptive order AR model, as demonstrated in this invention, show that the filtered residual sequence maintains a low noise baseline throughout the entire time period and clearly presents two high-amplitude isolated pulses. This is because the invention can dynamically adjust the order according to the working conditions, restore the geometric characteristics and amplitude intensity of the torsional impact signal, and effectively suppress environmental interference, achieving a high signal-to-noise ratio. This confirms that the invention can achieve a balance between suppressing background noise and preserving the fidelity of the torsional impact signal when dealing with variable downhole working conditions, demonstrating high adaptability and robustness.

Claims

1. A method for acquiring and processing torque impact signals from a PDC drill bit, characterized in that, include: Obtain the torque and rotation speed data of the PDC drill bit, and divide the torque data into several torque data windows; The torque waveform complexity of the torque data window is determined based on the mean of the numerical differences between adjacent data points within the torque data window and the standard deviation within the torque data window. Based on the rate of change of values ​​between adjacent data points in the speed data segment within the same time period corresponding to the torque data window, the speed slip strength of the torque data window is determined, and the formation meshing complexity of the torque data window is determined based on the torque waveform complexity and the speed slip strength. The adaptive order of the AR model for determining the torque data window is determined based on the formation meshing complexity. An adaptive order is used to perform predictive filtering on the torque data within the torque data window to obtain the predictive residual sequence of the torque data window, and the torque impact signal is extracted based on the numerical characteristics of the predictive residual sequence. The adaptive order of the AR model for the torque data window satisfies the expression: ; For the first Adaptive order of the AR model for a torque data window This is the maximum order of the preset AR model. The minimum order of the preset AR model. For the first Formation meshing complexity for each torque data window This is the floor function.

2. The method for acquiring and processing the torque impact signal of a PDC drill bit according to claim 1, characterized in that, The acquisition of torque and rotational speed data of the PDC drill bit includes: acquiring torque data of the downhole drill bit during rock cutting using a strain gauge torque sensor; and acquiring rotational speed data of the downhole drill bit during rock cutting using a MEMS gyroscope.

3. The method for acquiring and processing the torque impact signal of a PDC drill bit according to claim 1, characterized in that, The step of dividing the torque data into several torque data windows includes: dividing the torque data into several torque data windows of equal length using a sliding window.

4. The method for acquiring and processing the torque impact signal of a PDC drill bit according to claim 1, characterized in that, The torque waveform complexity of the torque data window satisfies the expression: ； In the formula, For the first The complexity of torque waveforms for each torque data window. The total number of data points within each torque data window. For the first The first torque data window Torque values ​​at each data point For the first The first torque data window Torque values ​​at each data point For the first The standard deviation of torque data within a single torque data window It is the maximum-minimum normalization function.

5. The method for acquiring and processing the torque impact signal of a PDC drill bit according to claim 1, characterized in that, The rotational slip strength of the torque data window satisfies the expression: ； In the formula, For the first The speed-slip strength of each torque data window The total number of data points within each torque data window. For the first The torque data window corresponds to the first speed data segment within the same time period. The rotational speed values ​​at each data point For the first The torque data window corresponds to the first speed data segment within the same time period. The rotational speed values ​​at each data point For the first Each torque data window corresponds to the sampling time interval of the speed data segment within the same time period. It is the maximum-minimum normalization function.

6. A method for acquiring and processing PDC drill bit torque impact signals according to claim 1 or 4, characterized in that, The formation engagement complexity of the torque data window satisfies the expression: ； In the formula, For the first Formation meshing complexity for each torque data window For the first The complexity of torque waveforms for each torque data window. For the first The speed-slip strength of each torque data window It is a natural exponential function.

7. The method for acquiring and processing the torque impact signal of a PDC drill bit according to claim 1, characterized in that, The method of predicting and filtering the torque data within the torque data window includes: using the Burg algorithm to estimate the autoregressive coefficients of the AR model of the torque data window based on the adaptive order of the AR model of the torque data window, and using the AR model of the torque data window to perform linear prediction on the torque data within the torque data window to obtain the predicted value of the torque data within the torque data window.

8. The method for acquiring and processing the torque impact signal of a PDC drill bit according to claim 7, characterized in that, The method for obtaining the prediction residual sequence of the torque data window is as follows: the difference between the torque data in the torque data window and the predicted value of the torque data in the torque data window is used as the prediction residual sequence of the torque data window.

9. The method for acquiring and processing the torque impact signal of a PDC drill bit according to claim 1, characterized in that, The extraction of the torsional impact signal includes: calculating the kurtosis value of the predicted residual sequence of the torsional data window; in response to the kurtosis value of the predicted residual sequence of the torsional data window being greater than a preset kurtosis threshold, a torsional impact signal exists in the torsional data window; and recording the time and amplitude of the torsional impact signal.

Images