Adaptive variable parameter drilling method for laminated structure based on interface confidence correction

Abstract

An adaptive variable-parameter drilling method for stacked structures based on interface confidence correction belongs to the field of machining technology. This method, based on the difference in machinability between adjacent material layers, constructs an interface proximity confidence model by weighted fusion of multi-source machining signals such as spindle current, feed current, axial force, torque, acoustic emission, vibration, and axial displacement, after obtaining the layer sequence, layer thickness, and material parameters. The theoretical variable-parameter region is dynamically corrected by advancing, delaying, or maintaining this model. This enables a smooth transition from soft to hard, hard to soft, and approximately equivalent interface transitions to the next layer's machining parameters within the corrected variable-parameter process window, following a continuous function. This reduces interface load impact and improves the stability of hole quality and tool life in stacked structures.

Description

Technical Field

[0001] This invention belongs to the field of machining technology and relates to an adaptive variable parameter drilling method for stacked structures based on interface confidence correction. Background Technology

[0002] Multi-layered structures are composed of various heterogeneous materials with significantly different physical and machinability properties. Typical structures include carbon fiber reinforced composite / metal laminates, composite / potting compound / metal laminates, and multiphase composite / titanium alloy laminates. These structures are characterized by lightweight, high structural strength, and good overall load-bearing capacity, and are widely used in aerospace, rail transportation, automotive manufacturing, and high-end equipment. To achieve the connection and assembly between laminated components and other parts, high-quality drilling is usually required. Due to significant differences in hardness, elastic modulus, thermal conductivity, fracture mode, chip morphology, and thermal sensitivity among the layers in a multi-layered structure, the appropriate spindle speed, feed rate, and cooling method control range differs for different materials. Using a single constant machining parameter for drilling often makes it difficult to simultaneously ensure the machining quality of each layer, easily leading to problems such as composite delamination and tearing, metal exit burrs, deterioration of hole wall roughness, composite ablation, rapid tool wear, and interface damage. Therefore, there is an urgent need to propose an adaptive variable parameter drilling method for the transition interface between soft and hard materials in multi-layered structures.

[0003] When drilling multi-layered structures, accurately identifying the variable parameter positions is fundamental to improving machining quality and tool life. Qin Xuda et al. from Tianjin University proposed an "Online Monitoring Adaptive Machining Method for Drilling Multi-Layered Structures" (patent application number CN201410504645.9). This method uses ultrasonic waves to measure the distance between the bottom of the drilling tool and the material interface in the multi-layered structure. The spindle speed and feed rate of the drilling tool are adjusted based on this distance. The parameter switching relies on preset interface determination conditions, rather than a comprehensive response to real-time cutting load, energy, or material removal status during drilling. The patent application number CN 202111228893.1, "An Online Interface Identification Method for Low-Frequency Vibration-Assisted Drilling of Multiphase Laminated Structures," proposed by Chen Yan et al. of Nanjing University of Aeronautics and Astronautics, describes a method that identifies the time it takes for the drill tip to reach the laminated interface during low-frequency vibration drilling of multiphase laminated structures. Although this method incorporates online signal analysis, its core function remains the identification of the interface "time point" and comparison with a preset threshold. It does not adaptively control the continuous load fluctuations caused by gradual changes in material properties or differences in hardness and softness within the interface transition zone. The patent application number CN202410261229.4, entitled "A Variable Parameter Machining Method Applicable to Deep Hole Drilling", proposed by Tang Zhijun et al. of CRRC Nanjing Puzhen Rolling Stock Co., Ltd., describes a method that reduces the feed rate by repeatedly drilling with variable parameters to decrease the frictional resistance when cutting into the workpiece, ensuring that the cutting edge does not slip when contacting the workpiece, thereby achieving deep hole drilling. The variable parameter process is mainly planned and controlled based on a preset displacement function relationship, without combining the actual cutting state, material interface changes, or real-time load information during the drilling process for closed-loop feedback adjustment.

[0004] The aforementioned studies and methods can address the interface location identification problem to some extent, but their focus is typically on "whether the interface has been reached," lacking clear control rules regarding "where parameter changes begin before and after the interface, how parameters change during the transition between soft and hard materials, and how to avoid load impacts during the switching process." Therefore, a multi-layered structure adaptive variable parameter drilling method is needed to comprehensively utilize prior material information, tool geometry, real-time machining signals, and control system response characteristics. This method would solve the problems of subjective parameter switching locations, insufficient adaptability of fixed thresholds, large impacts from sudden parameter changes, and unstable machining quality at the soft-hard material transition interface in existing methods. This would have significant engineering implications for improving the drilling quality of multi-layered structures and extending tool life. Summary of the Invention

[0005] This invention proposes an adaptive variable-parameter drilling method for stacked structures based on interface confidence correction. This method, based on the differences in machinability between adjacent layers, constructs an interface proximity confidence model by weighted fusion of multi-source machining signals, including spindle current, feed current, axial force, torque, acoustic emission, vibration, and axial displacement, after obtaining layer sequence, layer thickness, and material parameters. The theoretical variable-parameter region is dynamically corrected by advancing, delaying, or maintaining this model. This enables a smooth transition from soft to hard, hard to soft, and approximately equivalent interface transitions to the next layer's machining parameters within the corrected variable-parameter process window, following a continuous function. This reduces interface load impact and improves the stability of hole quality and tool life in stacked structures.

[0006] To achieve the above objectives, the present invention adopts the following technical solution.

[0007] An adaptive variable-parameter drilling method for stacked structures based on interface confidence correction includes the following steps: S1. Obtain the layer sequence, layer thickness, material type parameters, hardness parameters, thermal conductivity parameters, elastic modulus parameters, and fiber reinforcement characteristic parameters of each layer material in the multi-layer structure to be processed, as well as the corresponding recommended processing parameters. Normalize each parameter and establish a set of material processability parameters for each layer material. S2, establish the correspondence between material machinability parameters and machining parameters based on the set of material machinability parameters, forming a machining parameter database. The machining parameters include spindle speed n and cutting speed V. c Feed rate f, feed speed V f One or more of the following: axial force limit, current limit, torque limit, cooling method, and chip removal method; S3, calculate the processability difference parameter between adjacent layers based on the processability parameters of the two adjacent layers, and divide the interface between adjacent layers into soft material to hard material transition, hard material to soft material transition and approximately equivalent material transition based on the processability difference parameter between the two adjacent layers; S4, the theoretical variable parameter process window is determined based on the interface type of adjacent layers, the theoretical position of the interface, the tool diameter, the drill tip angle, the feed rate and the control response time; S5, during the drilling process, the machining signal is collected and the interface proximity confidence is calculated. The theoretical variable parameter process window is dynamically corrected based on the interface proximity confidence. S6, When the tool enters the modified variable parameter process window, the current layer machining parameters are smoothly transitioned to the next layer machining parameters according to a continuous function; S7, when the machining signal exceeds the allowable fluctuation range of the current material, execute one or more abnormal protection actions such as reducing feed rate, reducing spindle speed, pausing feed, retracting tool and removing chips, re-evaluating machining status, and alarm shutdown. S8, repeat S3 to S7 until the hole-making process of all layers of material is completed.

[0008] Furthermore, in S1: The normalization process includes: for parameters that are positively correlated with cutting difficulty, a forward normalization process is used; for parameters that are negatively correlated with cutting difficulty, a reverse normalization process is used, converting material characteristic parameters of different dimensions into dimensionless parameters in the range of 0 to 1.

[0009] The hardness parameters are obtained through a micro Vickers hardness tester, Rockwell hardness tester, Brinell hardness tester, material database, material certification documents, or historical processing data.

[0010] The set of material machinability parameters M[i] is obtained directly from the hardness range or by weighted calculation of normalized material characteristic parameters.

[0011] The normalized hardness parameter, material type coefficient, thermal conductivity parameter, elastic modulus parameter, and fiber reinforcement characteristic parameter are denoted as H′[i], T′[i], K′[i], E′[i], and F′[i], respectively, where i is the layer number of the material, with a value ≥ 1. The set of material processability parameters is obtained by weighted calculation of the normalized multi-type material characteristic parameters: Wherein, w1, w2, w3, w4, and w5 are the weighting coefficients corresponding to the hardness parameter, material type coefficient, thermal conductivity parameter, elastic modulus parameter, and fiber reinforcement characteristic parameter, respectively, and the sum of each weighting coefficient is 1. The weighting coefficients are determined based on machining experiment results, historical machining data, recommended tool parameters, or expert experience. The larger the value of M[i], the higher the equivalent cutting difficulty of the material layer, or the stronger the demand for low speed, small feed, and low load impact under the current tool and machining method. Through the above normalization process, material performance parameters of different dimensions and sources can be uniformly mapped to comparable material machinability parameters, providing a basis for subsequent determination of the soft-hard transition type of adjacent layers and calculation of variable parameter process windows.

[0012] Furthermore, in S2, the correspondence between the material machinability parameters and the machining parameters is obtained based on machining experiment results, numerical simulation results, tool manufacturer recommended parameters, and historical machining data.

[0013] Further, in S3: the machinability difference parameter between adjacent layers is: ΔMi=M[i+1]-M[i], where M[i] is the material machinability parameter of the i-th layer material, and M[i+1] is the material machinability parameter of the (i+1)-th layer material; when ΔMi>ε, it is determined to be a transition from soft material to hard material, that is, the cutting difficulty or equivalent machinability level of the lower layer material is higher than that of the upper layer material; when ΔMi<-ε, it is determined to be a transition from hard material to soft material, that is, the cutting difficulty or equivalent machinability level of the upper layer material is higher than that of the lower layer material; when |ΔMi|≤ε, it is determined to be a transition of approximately equivalent materials, and the current parameter can be maintained or a small correction can be made, where ε is a preset difference threshold.

[0014] Furthermore, in S4: when the interface between adjacent layers is a transition from soft to hard material, the theoretical variable-parameter process window is set before the theoretical position of the interface. The starting point Zs and the ending point Ze of the window satisfy: Zs = Zi - δi, Ze = Zi, where δi > 0. Here, Zi is the theoretical position of the interface between adjacent layers, and δi is the length of the theoretical variable-parameter process window. This allows the machining parameters to be gradually adjusted to those suitable for hard materials before the tool tip actually contacts the hard material, preventing the tool tip from suddenly entering the hard material under soft material parameters. The effect of this strategy is to gradually adjust the spindle speed, feed rate, or feed speed to suitable machining parameters for hard materials before the tool tip actually contacts the hard material, ensuring that the tool is in a low-impact, stable cutting state when entering the hard layer. This reduces the instantaneous axial force and torque mutation at the soft-hard interface, avoiding tool tip chipping, edge chipping, hard layer inlet hole wall scratches, and thermal damage at the interface caused by excessive feed or mismatched cutting speeds, while improving the machining stability during the hard material entry stage.

[0015] When the interface between adjacent layers is a transition from hard to soft material, to avoid prematurely switching to soft material parameters while the main cutting edge is still cutting within the hard material, the theoretical variable parameter process window is set after the main cutting edge of the tool has completely detached from the upper hard material. The window start point Zs and the window end point Ze satisfy: Zs = Zi + Lc, Ze = Zi + Lc + δi 、 Where Lc is the axial distance required for the main cutting edge of the tool to completely cross the interface, and the theoretical variable parameter process window length δi is at this point. 、 The calculation formula is: Where τ is the control response time and δ0 is the safety correction amount.

[0016] Under conditions with significant material differences, δi 、 Corrections should be made based on material differences. In the formula, k M This is the material difference correction factor.

[0017] The formula for calculating Lc is: Where D is the cutter diameter and κ is the drill bit tip angle.

[0018] The purpose of this strategy is to maintain low speed, small feed, or low load parameters suitable for hard materials until the hard material is completely removed. This prevents chipping, increased burrs, hole wall step damage, and abnormal tool wear at the hard layer exit when the main cutting edge has a hard layer remaining, which would occur if high feed or high efficiency parameters for soft materials were used prematurely. After the main cutting edge has completely removed the hard layer, the machining parameters are smoothly transitioned to those for soft materials. This ensures the quality of the hard layer exit while improving the machining efficiency of subsequent soft material, and reduces the impact of abrupt parameter changes in the interface transition zone.

[0019] When the interface between adjacent layers is a transition of approximately equivalent materials, a complete parameter switch is not performed, or only one or more of the spindle speed, feed rate, and feed speed are slightly modified, and the amount of the slight modification does not exceed the preset change threshold of the corresponding machining parameter.

[0020] Further, in S5: the machining signal includes one or more of the following: spindle current, feed current, axial force, torque, acoustic emission signal, vibration signal, and spindle power; the interface proximity confidence is obtained by weighted fusion of one or more of the following confidence values: Z-axis theoretical distance feature (extracted from Z-axis displacement information fed back by the CNC system), current change feature, axial force change feature, torque change feature, acoustic emission frequency feature, and vibration amplitude change feature. The formula is: Where Cz is the confidence level of the theoretical distance along the Z-axis, Ci is the confidence level of the current variation characteristic, Cf is the confidence level of the axial force variation characteristic, Ct is the confidence level of the torque variation characteristic, Cae is the confidence level of the acoustic emission characteristic, and Wz, Wi, Wf, Wt, and Wae are the corresponding weights, with the sum of each weight being 1. Based on actual processing conditions, the confidence levels of features with small variation ranges for different materials can be ignored, while ensuring that at least one weight is not zero. When C is greater than or equal to the preset threshold Cth and remains continuously for m sampling cycles, it is determined that the tool has entered the variable parameter process window or interface transition state.

[0021] Dynamic correction of the theoretical variable parameter process window includes: when the theoretical Z-axis position is consistent with the interface proximity confidence judgment result, the variable parameter process window is executed according to the theoretical variable parameter process window; when there is a deviation between the theoretical Z-axis position and the interface proximity confidence, Zs and Ze are corrected according to the interface proximity confidence C; when the machining signal shows a trend of matching the next layer material ahead of time, the variable parameter process window is corrected forward; when the machining signal still maintains the stable cutting characteristics of the current layer, the variable parameter process window is corrected backward.

[0022] Further, in S6: the continuous function is a linear function, a piecewise linear function, a sigmoid function, or a stepwise gradually varying function; the processing parameter P(z) is determined according to... The transition is performed, where Pi is any current layer processing parameter, P{i+1} is the next layer processing parameter, and η(z) is a transition function between 0 and 1. For example, when the current variable parameter progress S = (Z-Zs) / (Ze-Zs), the sigmoid function can be written as: By continuously varying parameters, the impact of cutting load caused by sudden changes in rotational speed or feed rate can be avoided, thereby improving the hole wall quality at the interface and the working stability of the tool.

[0023] Furthermore, in S8: for the last layer of material, the corresponding processing parameters for that layer of material are used to continuously drill until the material is completely drilled through, and the tool retraction and chip removal actions are performed according to the monitoring signal.

[0024] The beneficial effects of this invention are as follows: This invention no longer uses "interface arrival" as the sole trigger for parameter variation. Instead, based on the difference in machinability between adjacent material layers, it divides the interface transition process into three categories: soft material to hard material, hard material to soft material, and approximately equivalent material. Furthermore, it introduces geometric parameters such as tool diameter and drill tip angle, combined with the actual participation state of the tool tip and main cutting edge relative to the material interface, to determine the parameter variation position for early, delayed, or stable transition. This ensures that the parameter variation strategy matches the material transition characteristics and the tool's geometric cutting state, avoiding premature switching when the hard layer is not fully machined or delayed switching when the hard layer enters the cut. Simultaneously, this invention, through the main... Multi-source machining signals, including shaft current, feed current, axial force, torque, and acoustic emission, are used to construct an interface with near-confidence levels. This allows for dynamic correction of the theoretical layer thickness and theoretical interface position. Within a variable-parameter process window, continuous functions are used to smoothly adjust machining parameters such as spindle speed and feed rate, thereby reducing the impact of cutting loads caused by parameter abrupt changes and suppressing composite material delamination and tearing, metal exit burrs, composite material ablation, and interface hole wall damage. This method can be integrated with CNC machine tools, robotic hole-making equipment, automatic feed drills, portable hole-making systems, and low-frequency vibration-assisted hole-making systems. It features clear control logic, wide applicability, and strong engineering application adaptability. Attached Figure Description

[0025] Figure 1 Flowchart of a multi-layer adaptive variable parameter hole-making method; Figure 2 A schematic diagram of the pre-parameter variable process window for the transition interface between soft and hard materials; Figure 3 A schematic diagram of the delayed variable parameter process window for the transition interface between hard and soft materials; Figure 4 This is a schematic diagram of the multi-source processing signal fusion and continuous parameter variable control logic. Detailed Implementation

[0026] The present invention will be further described below with reference to the accompanying drawings and specific embodiments. The following embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. All equivalent substitutions, modifications, or improvements made based on the inventive concept should be included within the scope of protection of the present invention.

[0027] Example 1: A multi-layered structure composed of T800 grade carbon fiber reinforced resin matrix composite material / potting compound / T800 grade carbon fiber reinforced resin matrix composite material / TC4 titanium alloy was used as the processing object. The first layer of T800 grade carbon fiber reinforced resin matrix composite material had a thickness of 6 mm, the second layer of potting compound had a thickness of 40 mm, the third layer of T800 grade carbon fiber reinforced resin matrix composite material had a thickness of 6 mm, and the fourth layer of TC4 titanium alloy had a thickness of 5 mm. A twist drill with a diameter D=8 mm and a drill bit tip angle κ=130° was used for vertical hole drilling, and a dust extraction or chip removal device was used in conjunction.

[0028] In this embodiment, the hardness of each layer of material and the corresponding processing parameters are shown in Table 1.

[0029] Table 1. Machinability parameters and recommended machining parameters for each layer of the multi-layer structure. When the first layer of T800 grade CFRP enters the second layer of potting compound, ΔM1 < -ε, indicating a transition from hard to soft material. To avoid prematurely switching to the high feed parameters of the potting compound while the main cutting edge is still cutting within the CFRP, the parameter transition should only be initiated after the main cutting edge has completely detached from the first layer of CFRP. From D = 8 mm and κ = 130°, we can obtain: (11) Therefore, after the drill tip reaches the theoretical interface between the first and second layers, continue drilling downwards for about 1.87 mm, or after confirming through multi-source signals that the main cutting edge has detached from the first layer, smoothly transition the machining parameters from B to A.

[0030] When the second layer of potting compound enters the third layer of T800 grade CFRP, ΔM2 > ε, indicating a transition from soft to hard material. To prevent the tool tip from suddenly entering the CFRP under the potting compound parameters, a parameter transition is initiated before the drill tip reaches the theoretical interface between the second and third layers. The advance distance δ2 is determined based on the feed rate, control response time, and safety correction. When δ2 is 0.5 mm, the variable parameter window is set to Zs = Zi - 0.5 mm, Ze = Zi, and the machining parameters are smoothly transitioned from A to B within this window.

[0031] When the third layer of T800 grade CFRP transitions into the fourth layer of TC4 titanium alloy, ΔM3 > ε, indicating a transition from soft to hard material. Since the hardness and cutting resistance of TC4 titanium alloy are significantly higher than CFRP, the variable parameter window can be activated before the theoretical interface and can be further advanced or delayed based on the current increase trend, axial force slope change, and acoustic emission frequency shift. The control system smoothly transitions the machining parameters from B to E within the variable parameter window, preventing the tool from suddenly entering the titanium alloy layer at high speeds or large feed parameters.

[0032] Example 2: Based on Example 1, the control system uses spindle current, axial force, and acoustic emission signals as multi-source machining signals. The root mean square value and rate of change are extracted from the spindle current signal, the slope and peak value changes are extracted from the axial force signal, and the dominant frequency offset is extracted from the acoustic emission signal. After normalization of each signal characteristic, the input interface approximates the confidence model. When C ≥ Cth and remains continuously for m sampling cycles, the tool is determined to have entered the interface transition state, and the variable parameter window is activated or corrected.

[0033] Example 3: During continuous hole drilling, when the spindle current exceeds the current material current limit, or the axial force exceeds the current material axial force limit, or an abnormally high amplitude sudden change occurs in the acoustic emission signal, the control system determines that there is a risk of tool overload, poor chip removal, or interface damage in the current hole, and executes actions such as reducing the feed rate, reducing the spindle speed, pausing the feed, short-distance tool retraction for chip removal, or alarm shutdown. Through the above-mentioned abnormal protection mechanism, the risk of tool breakage and loss of hole quality control can be reduced.

[0034] The above embodiments demonstrate that the present invention can combine the direction of material machinability differences, tool geometry participation state, and dynamic feedback of multi-source machining signals to form an adaptive variable parameter hole control method suitable for multi-layered structures. Compared with variable parameter methods based solely on interface arrival or fixed thresholds, the present invention can more rationally determine the timing of parameter changes and reduce interface machining impact through continuous parameter changes, thereby improving hole quality consistency and tool life.

Claims

1. An adaptive variable-parameter drilling method for stacked structures based on interface confidence correction, characterized in that, Includes the following steps: S1. Obtain the layer sequence, layer thickness, material type parameters, hardness parameters, thermal conductivity parameters, elastic modulus parameters, and fiber reinforcement characteristic parameters of each layer material in the multi-layer structure to be processed, as well as the corresponding recommended processing parameters. Normalize each parameter and establish a set of material processability parameters for each layer material. S2, establish the correspondence between material machinability parameters and machining parameters based on the set of material machinability parameters, forming a machining parameter database. The machining parameters include spindle speed n and cutting speed V. c Feed rate f, feed speed V f One or more of the following: axial force limit, current limit, torque limit, cooling method, and chip removal method; S3, calculate the processability difference parameter between adjacent layers based on the processability parameters of the two adjacent layers, and divide the interface between adjacent layers into soft material to hard material transition, hard material to soft material transition and approximately equivalent material transition based on the processability difference parameter between the two adjacent layers; S4, the theoretical variable parameter process window is determined based on the interface type of adjacent layers, the theoretical position of the interface, the tool diameter, the drill tip angle, the feed rate and the control response time; S5, during the drilling process, the machining signal is collected and the interface proximity confidence is calculated. The theoretical variable parameter process window is dynamically corrected based on the interface proximity confidence. S6, When the tool enters the modified variable parameter process window, the current layer machining parameters are smoothly transitioned to the next layer machining parameters according to a continuous function; S7, when the machining signal exceeds the allowable fluctuation range of the current material, execute one or more abnormal protection actions such as reducing feed rate, reducing spindle speed, pausing feed, retracting tool and removing chips, re-evaluating machining status, and alarm shutdown. S8, repeat S3 to S7 until the hole-making process of all layers of material is completed.

2. The adaptive variable-parameter drilling method for stacked structures based on interface confidence correction according to claim 1, characterized in that, In S1: The normalization process includes: for parameters that are positively correlated with cutting difficulty, a forward normalization process is used; for parameters that are negatively correlated with cutting difficulty, a reverse normalization process is used, converting material characteristic parameters of different dimensions into dimensionless parameters in the range of 0 to 1. The hardness parameters mentioned above are obtained through a micro Vickers hardness tester, Rockwell hardness tester, Brinell hardness tester, material database, material certification documents, or historical processing data; The set of material machinability parameters M[i] is obtained directly from the hardness range or by weighted calculation of normalized material characteristic parameters; The normalized hardness parameter, material type coefficient, thermal conductivity parameter, elastic modulus parameter, and fiber reinforcement characteristic parameter are denoted as H′[i], T′[i], K′[i], E′[i], and F′[i], respectively, where i is the layer number of the material, with a value ≥ 1; the set of material processability parameters is obtained by weighted calculation of the normalized multi-type material characteristic parameters: Among them, w1, w2, w3, w4 and w5 are the weighting coefficients corresponding to the hardness parameter, material type coefficient, thermal conductivity parameter, elastic modulus parameter and fiber reinforcement characteristic parameter, respectively, and the sum of each weighting coefficient is 1; the weighting coefficients are determined based on the processing test results, historical processing data, tool recommended parameters or expert experience; the larger the value of M[i], the higher the equivalent cutting difficulty of the material layer, or the stronger the demand for low speed, small feed and low load impact under the current tool and processing method.

3. The adaptive variable-parameter drilling method for stacked structures based on interface confidence correction according to claim 2, characterized in that, In S2, the correspondence between material machinability parameters and machining parameters is obtained based on machining experiment results, numerical simulation results, tool manufacturer recommended parameters, and historical machining data.

4. The adaptive variable-parameter drilling method for stacked structures based on interface confidence correction according to claim 3, characterized in that, In S3: the processability difference parameter between adjacent layers is: ΔMi=M[i+1]-M[i], where M[i] is the processability parameter of the i-th layer material and M[i+1] is the processability parameter of the (i+1)-th layer material; when ΔMi>ε, it is determined to be a transition from soft material to hard material; when ΔMi<-ε, it is determined to be a transition from hard material to soft material; when |ΔMi|≤ε, it is determined to be a transition of approximately equivalent materials, where ε is a preset difference threshold.

5. The adaptive variable-parameter drilling method for stacked structures based on interface confidence correction according to claim 4, characterized in that, In S4: when the interface between adjacent layers is a transition from soft material to hard material, the theoretical variable parameter process window is set before the theoretical position of the interface. The starting point Zs and the ending point Ze of the window satisfy: Zs=Zi-δi, Ze=Zi, and at this time δi>0, where Zi is the theoretical position of the interface between adjacent layers and δi is the length of the theoretical variable parameter process window. When the interface between adjacent layers is a transition from hard material to soft material, the theoretical variable parameter process window is set after the main cutting edge of the tool has completely detached from the upper hard material. The window start point Zs and the window end point Ze satisfy: Zs = Zi + Lc, Ze = Zi + Lc + δi 、 Where Lc is the axial distance required for the main cutting edge of the tool to completely cross the interface, and the theoretical variable parameter process window length δi is at this point. 、 The calculation formula is: Where τ is the control response time and δ0 is the safety correction amount; Under conditions with significant material differences, δi 、 Corrections should be made based on material differences. In the formula, k M This is a correction factor for material differences; The formula for calculating Lc is: Where D is the cutter diameter and κ is the drill bit tip angle; When the interface between adjacent layers is a transition of approximately equivalent materials, a complete parameter switch is not performed, or only one or more of the spindle speed, feed rate, and feed speed are slightly modified, and the amount of the slight modification does not exceed the preset change threshold of the corresponding machining parameter.

6. The adaptive variable-parameter drilling method for stacked structures based on interface confidence correction according to claim 5, characterized in that, In S5: the machining signal includes one or more of the following: spindle current, feed current, axial force, torque, acoustic emission signal, vibration signal, and spindle power; the interface proximity confidence level is obtained by weighted fusion of one or more of the following confidence levels: Z-axis theoretical distance characteristic, current change characteristic, axial force change characteristic, torque change characteristic, acoustic emission frequency characteristic, and vibration amplitude change characteristic; the formula is: Where Cz is the confidence level of the theoretical distance along the Z-axis, Ci is the confidence level of the current change characteristic, Cf is the confidence level of the axial force change characteristic, Ct is the confidence level of the torque change characteristic, Cae is the confidence level of the acoustic emission characteristic, and Wz, Wi, Wf, Wt and Wae are the corresponding weights, and the sum of each weight is 1; when C is greater than or equal to the preset threshold Cth and is maintained continuously for m sampling cycles, it is determined that the tool has entered the variable parameter process window or interface transition state. Dynamic correction of the theoretical variable parameter process window includes: when the theoretical Z-axis position is consistent with the interface proximity confidence judgment result, the variable parameter process window is executed according to the theoretical variable parameter process window; when there is a deviation between the theoretical Z-axis position and the interface proximity confidence, Zs and Ze are corrected according to the interface proximity confidence C; when the machining signal shows a trend of matching the next layer material ahead of time, the variable parameter process window is corrected forward; when the machining signal still maintains the stable cutting characteristics of the current layer, the variable parameter process window is corrected backward.

7. The adaptive variable-parameter drilling method for stacked structures based on interface confidence correction according to claim 6, characterized in that, In S6: the continuous function is a linear function, a piecewise linear function, a sigmoid function, or a stepwise gradually varying function; the processing parameter P(z) is according to... Transition is performed, where Pi is any current layer processing parameter, P{i+1} is the next layer processing parameter, and η(z) is a transition function between 0 and 1.

8. The adaptive variable-parameter drilling method for stacked structures based on interface confidence correction according to claim 7, characterized in that, In S8: For the last layer of material, the corresponding processing parameters for that layer of material are used to continuously drill until the material is completely drilled through, and the tool retraction and chip removal actions are performed according to the monitoring signal.

Images