One-step machining method for head high precision stepped deep hole

Abstract

This invention relates to the field of cutting technology and proposes a one-stop machining method for high-precision stepped deep holes with a head-to-head joint. The method includes: recording the initial temperature of several key thermally sensitive points after machine tool preheating; acquiring the left and right spindle torque signals and the temperature of each key thermally sensitive point during the current workpiece machining process; generating real-time tool deflection compensation amounts based on the real-time cutting force obtained at each moment during the machining process; acquiring real-time thermal drift compensation amounts at each moment; obtaining the total radial compensation amount at each moment; obtaining the approach strength at each moment; obtaining the instability measurement coefficient at each moment; obtaining the approach strength compensation weight at each moment; obtaining the final compensation amount at each moment and outputting the command to the machine tool CNC system; and superimposing the theoretical cutting trajectory to adjust the tool tip position. This invention aims to solve the problem of tool drift caused by elastic deformation and machine tool thermal expansion affecting accuracy during head-to-head machining.

Description

Technical Field

[0001] This invention relates to the field of cutting and machining technology, specifically to a one-stop machining method for high-precision stepped deep holes with a mating head. Background Technology

[0002] Stepped deep bores refer to precision internal cavity structures with multiple coaxial inner diameter sections and a length-to-diameter ratio significantly larger than that of conventional bores. Such parts are commonly found in high-end hydraulic integrated valve bodies, aerospace actuators, and ultra-long precision spindles. Their core quality requirements are extremely high, including strict dimensional tolerances, excellent geometric accuracy, and extremely high coaxiality between each stepped bore section.

[0003] To address the challenge of axial misalignment in high-precision machining of deep, elongated stepped holes caused by the coupling of multiple factors such as unpredictable internal conditions, dynamic tool deflection, and thermal drift, existing technologies primarily rely on offline measurement, post-processing correction, or operator experience. They lack a one-stop solution capable of real-time sensing of the docking status and simultaneous dynamic compensation during machining. Due to the relative position and orientation of the tools within the workpiece, it is difficult to predictively compensate for elastic deformation of the tools caused by changes in cutting force and systematic drift due to machine tool thermal expansion. This makes the final docking accuracy heavily dependent on the initial geometric accuracy and stable operating conditions of the machine tool. When dealing with material property fluctuations, complex stepped cross-sections, and long-term machining heat accumulation, reliability is poor, easily leading to quality defects such as docking misalignment and vibration marks at the steps, resulting in low machining consistency and success rate. Summary of the Invention

[0004] This invention provides a one-stop machining method for high-precision stepped deep holes with mating ends, to solve the problem of accuracy being affected by the drift of the cutting tools on both sides due to elastic deformation and thermal expansion of the machine tool in existing mating machining. The specific technical solution adopted is as follows:

[0005] This invention proposes a one-stop machining method for high-precision stepped deep holes with a mating head, the method comprising the following steps:

[0006] Record the initial temperature of several key thermally sensitive points after the machine tool is preheated; acquire the left and right spindle torque signals and the temperature of each key thermally sensitive point during the current workpiece machining process;

[0007] Based on the cutting cross-sectional area and actual feed rate obtained at each moment during the machining process, the real-time material removal rate at each moment is calculated. Combined with the cutting force coefficient per unit removal rate obtained from historical machining data, the real-time cutting force at each moment is obtained, and then the real-time tool deflection compensation amount at each moment is generated. Based on the temperature change at each moment and the difference from the initial temperature, and combined with the radial thermal sensitivity coefficient measured by the machine tool preheating, the real-time thermal drift compensation amount at each moment is obtained. Combined with the real-time tool deflection compensation amount, the total radial compensation amount at each moment is obtained.

[0008] The maximum cross-correlation coefficient of the left and right spindle torque signals at each moment is obtained based on the sliding window, and then the approach intensity at each moment is obtained by combining the cross-correlation coefficient changes during the machining process; the peak phase lag of the cross-correlation is obtained and combined with threshold analysis to obtain the axis skew coefficient at each moment; the approach intensity is low-pass filtered to obtain the local trend estimate at each moment, and then the trend deviation at each moment is obtained and combined with the axis skew coefficient to obtain the instability measurement coefficient at each moment; the approach intensity compensation weight at each moment is obtained by combining the approach intensity.

[0009] Based on the total radial compensation amount and the approach strength compensation weight, the final compensation amount at each moment is obtained and the command is output to the machine tool CNC system. The theoretical cutting trajectory is superimposed to adjust the tool tip position. Based on the adjusted cutting state, a control cycle is formed until the docking machining is completed.

[0010] Optionally, the real-time material removal rate at each moment is calculated, and the cutting force coefficient per unit removal rate obtained through historical processing data is used to obtain the real-time cutting force at each moment. The specific method includes:

[0011] The product of the cutting cross-sectional area and the actual feed rate at any given moment is taken as the real-time material removal rate at that moment.

[0012] Based on the cutting force of the same workpiece material and tool type as the current workpiece under different material removal rates in historical processing data, the cutting force coefficient per unit removal rate is obtained by linear regression fitting.

[0013] The product of the real-time material removal rate at that moment and the cutting force coefficient per unit removal rate is taken as the real-time cutting force at that moment.

[0014] Optionally, the real-time tool deflection compensation amount at each moment is obtained using the following method:

[0015] The deflection coefficient is calculated based on the tool overhang length, elastic modulus, and moment of inertia of the cross section.

[0016] The product of the real-time cutting force at any given moment and the deflection coefficient is taken as the real-time tool deflection compensation amount at that moment.

[0017] Optionally, the real-time thermal drift compensation amount at each moment is obtained using the following method:

[0018] The difference between the temperature of any critical thermally sensitive point at any given time and the initial temperature is taken as the real-time temperature rise of that critical thermally sensitive point at that time; the average of the real-time temperature rises of all critical thermally sensitive points at that time is taken as the real-time temperature rise at that time.

[0019] The radial thermal sensitivity coefficient is obtained based on the radial displacement change of the standard test bar during the machine tool preheating process and the temperature rise change during the machine tool preheating process.

[0020] The product of the real-time temperature rise at that moment and the radial thermal sensitivity coefficient is used as the real-time thermal drift compensation amount at that moment.

[0021] Optionally, the specific method for obtaining the total radial compensation at each time point includes:

[0022] The sum of the real-time tool deflection compensation and the real-time thermal drift compensation at any moment on any side of the spindle is taken as the total radial compensation of that side of the spindle at that moment.

[0023] Optionally, the method for obtaining the maximum cross-correlation coefficient of the left and right spindle torque signals at each moment based on a sliding window, and then combining the cross-correlation coefficient changes during the machining process to obtain the approach intensity at each moment, includes the following specific methods:

[0024] For any moment in the two torque signals of the left and right spindles, a sliding window is used to obtain a segment of local signal corresponding to each of the two torque signals of the left and right spindles at that moment. The time delay variable and its value range are preset, and the cross-correlation coefficient between the two segments of local signal under different time delay variables is obtained. The maximum value of the cross-correlation coefficient under all time delay variables is taken as the maximum value of the cross-correlation coefficient at that moment.

[0025] Based on the mean and standard deviation of the maximum cross-correlation coefficients at each time point, the cross-correlation coefficient threshold is obtained; according to the time sequence, the first time point among the two times when the maximum cross-correlation coefficients of two consecutive times are greater than or equal to the cross-correlation coefficient threshold is taken as the background cutoff time; the mean of the maximum cross-correlation coefficients of all times from the first time point in the processing to the time point before the background cutoff time is taken as the background correlation level.

[0026] For any given moment, the difference between the maximum cross-correlation coefficient at that moment and the background correlation level is taken as the additional coupling strength at that moment; the difference between 1 and the background correlation level is taken as the background correlation removal level; and the ratio of the additional coupling strength at that moment to the background correlation removal level is taken as the approach strength at that moment.

[0027] Optionally, the specific method for obtaining the axis skew coefficient at each time point includes:

[0028] Based on the maximum cross-correlation coefficient of the left and right spindle torque signals at each moment, obtain the maximum cross-correlation coefficient at each moment and the phase lag corresponding to the peak value of the maximum cross-correlation coefficient;

[0029] A maximum permissible phase lag threshold is preset, and the ratio of the absolute value of the phase lag at any given moment to the maximum permissible phase lag threshold is used as the axis skew coefficient at that moment.

[0030] Optionally, the specific method for obtaining the instability metric coefficients at each time point includes:

[0031] The absolute value of the difference between the approach strength at any given time and the local trend estimate at that time is taken as the trend deviation at that time.

[0032] The product of the ratio of the trend deviation at that moment to the local trend estimate and the axis skewness coefficient at that moment is used as the instability measure coefficient at that moment.

[0033] Optionally, the specific method for obtaining the approach intensity compensation weights at each time point includes:

[0034] Using the approach intensity at any given time as the base, an exponential function is constructed by adding 1 to the instability metric coefficient at that time. The approach intensity compensation weight at that time is obtained based on the exponential function, and the approach intensity compensation weight is negatively correlated with the output value of the exponential function.

[0035] Optionally, the specific method for obtaining the final compensation amount at each time point includes:

[0036] The product of the total radial compensation amount of the spindle on one side at any time and its approach strength compensation weight is taken as the final compensation amount of the spindle on that side at that time, and the final compensation amount of the spindle on the other side at that time is obtained.

[0037] The beneficial effects of this invention are as follows: By constructing an online closed-loop control system based on multi-source signal fusion and fuzzy state perception, this invention effectively overcomes the technical bottleneck of tool elastic deformation caused by cutting force changes and systemic drift caused by machine tool thermal expansion affecting docking accuracy. Through cross-correlation analysis of the left and right spindle torque signals, it constructs two fuzzy quantitative indicators: approach strength and axis skew coefficient, achieving non-contact, real-time perception of the internal docking state. Simultaneously, it obtains real-time cutting force by analyzing material removal rate, quantifies tool deflection compensation by combining tool deflection, and analyzes temperature changes and thermal drift to obtain thermal drift compensation. This system achieves online feedforward compensation for systematic errors, reducing trajectory deviation at its source. Based on the approach strength and axis skew coefficient, a dynamic compensation weight algorithm based on process stability is designed to intelligently adjust the compensation amount to quantify the approach strength compensation weight, avoiding the risk of overcompensation in the force chain coupling stage. Finally, by integrating real-time state perception and multi-physics compensation, an adaptive closed-loop control command is formed to directly drive the machine tool for accuracy correction, thereby significantly improving the coaxiality and success rate of head docking, enhancing the adaptability and stability of the machining process to complex working conditions, and realizing one-stop high-quality machining of high-precision stepped deep holes. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0039] Figure 1 This is a schematic diagram of a one-stop machining method for high-precision stepped deep holes provided in an embodiment of the present invention. Detailed Implementation

[0040] 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 embodiments of the present invention, and not all embodiments. 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.

[0041] Please see Figure 1 The diagram illustrates a one-stop machining method for high-precision stepped deep holes according to an embodiment of the present invention. The method includes the following steps:

[0042] Step S001: Record the initial temperature of several key thermally sensitive points after the machine tool is preheated; acquire the left and right spindle torque signals and the temperature of each key thermally sensitive point during the current workpiece machining process.

[0043] The purpose of this embodiment is to address the issue that during head-to-head machining, two cutting tools may align with the target center in opposite directions, potentially leading to misalignment. Therefore, the aim is to compensate for errors during the alignment process. This requires acquiring system references and real-time acquisition of multi-channel synchronous processing signals before and during machining.

[0044] Specifically, before machining, a spindle no-load operation program is executed, with the speed covering the commonly used speed range during machining; this embodiment does not impose specific limitations. Torque and current signals of the left and right spindle shaft drives are collected, with the sampling frequency set to 2000Hz. The background noise level is calculated based on the variance of the no-load current signal. After the machine tool is preheated for 30 minutes, the temperature of several key thermally sensitive points in the initial state is recorded and used as the initial temperature of each key thermally sensitive point. The key thermally sensitive points include the spindle box bearing seat, the column guide rail mating surface, the area near the bed anchor bolts, and the feed screw nut seat, etc., with a high-precision temperature sensor arranged at each location.

[0045] Furthermore, based on the unified clock signal of the machine tool CNC system, the torque signals of the left and right spindles, the feed axis load current, and the temperature of each key thermally sensitive point at each moment are synchronously collected during the machining process. Gaussian filters are applied to the torque and current signals for noise reduction, and the no-load torque signal is subtracted from the torque signal to remove the influence of the machine tool itself. The final processed torque signal is used as the torque signal of the left and right spindles in subsequent processing. The Gaussian filter noise reduction and the removal of the no-load torque signal from the torque signal are existing technologies and will not be described in detail in this embodiment.

[0046] Step S002: Based on the cutting cross-sectional area and actual feed rate obtained at each moment during the machining process, the real-time material removal rate at each moment is calculated. Combined with the cutting force coefficient per unit removal rate obtained through historical machining data, the real-time cutting force at each moment is obtained, and then the real-time tool deflection compensation amount at each moment is generated. Based on the temperature change at each moment and the difference from the initial temperature, and combined with the radial thermal sensitivity coefficient measured by the machine tool preheating, the real-time thermal drift compensation amount at each moment is obtained. Combined with the real-time tool deflection compensation amount, the total radial compensation amount at each moment is obtained.

[0047] It should be noted that in deep hole machining, the tool overhang is large, and the cutting force will cause the tool to deflect, resulting in the tool tip deviating from the theoretical trajectory. At the same time, the temperature of key parts such as the spindle box and column will rise during the operation of the machine tool, which will cause thermal expansion, further aggravating the positional deviation, thus causing the axis to misalign during docking.

[0048] Preferably, in one embodiment of the present invention, based on the cutting cross-sectional area and actual feed rate obtained at each moment during the machining process, the real-time material removal rate at each moment is calculated. Combined with the cutting force coefficient per unit removal rate obtained through historical machining data, the real-time cutting force at each moment is obtained, and then the real-time tool deflection compensation amount at each moment is generated. The specific method includes:

[0049] It should be noted that in the relevant technologies in this field, during the finishing stage of deep hole machining where the cutting speed (spindle speed) remains constant, the principal component of the cutting force (tangential force) is proportional to the "cutting volume per revolution", that is, the cutting volume per revolution = cutting cross-sectional area × feed per revolution.

[0050] Specifically, for any given moment, based on the machining diameter and step position at that moment during the machining process, the cross-sectional area that the tool is cutting at that moment is obtained and recorded as the cutting cross-sectional area at that moment. At the same time, the actual feed rate at that moment is read from the feed axis servo driver, and the product of the cutting cross-sectional area at that moment and the actual feed rate is used as the real-time material removal rate at that moment.

[0051] It should be noted that a higher real-time material removal rate indicates a heavier processing load, a larger expected cutting force, and more severe tool deflection, requiring a larger compensation amount. However, directly measuring the cutting force requires an expensive force gauge and is difficult to install, making it unsuitable for industrial sites. When a long overhanging tool is under force, it can be simplified into a cantilever beam model because tool deflection is caused by the cutting force, and the cutting force is proportional to the material removal rate, so the amount of deflection is also proportional to the cutting force.

[0052] Furthermore, based on historical processing data, the cutting force of the same workpiece material and tool type under different material removal rates is obtained. That is, cutting tests are conducted on the same workpiece material and tool type under different material removal rates, and the corresponding cutting force is measured. Based on the material removal rate and the corresponding cutting force, the cutting force coefficient per unit removal rate is obtained by linear regression fitting. The cutting force coefficient per unit removal rate is the slope of the straight line obtained by linear regression fitting, where the horizontal axis is the material removal rate and the vertical axis is the cutting force. Linear regression fitting is a prior art and will not be described in detail in this embodiment.

[0053] It should be noted that the larger the cutting force coefficient value per unit removal rate, the harder the material and the duller the tool, reflecting the ease with which the workpiece material is removed and the sharpness of the tool.

[0054] Furthermore, the product of the real-time material removal rate at any given moment and the cutting force coefficient per unit removal rate is taken as the real-time cutting force at that moment.

[0055] Furthermore, in mechanics of materials, the deflection at the tool tip... The tool overhang length (effective overhang length of the tool as the machining depth changes), the elastic modulus of the tool material, and the moment of inertia of the tool section can all be obtained a priori. The deflection coefficient is then calculated based on the tool overhang length, elastic modulus, and moment of inertia of the section, simplifying the calculation of the constant term to the deflection coefficient. The product of the real-time cutting force at any given moment and the deflection coefficient is taken as the real-time tool deflection compensation at that moment. The real-time tool deflection compensation is a length quantity in micrometers, representing the expected radial offset of the tool tip due to the cutting force under the real-time material removal rate.

[0056] It should be further noted that solid materials undergo thermal expansion / contraction when the temperature changes. The greater the temperature change, the greater the change in length. At the same time, different materials have different thermal expansion capabilities, and the absolute expansion of an object is related to its original length. Therefore, the change in length is directly proportional to the material's coefficient of thermal expansion and its original length.

[0057] Preferably, in one embodiment of the present invention, the real-time thermal drift compensation amount at each moment is obtained based on the difference between the temperature change at each moment and the initial temperature, and in combination with the radial thermal sensitivity coefficient measured during machine tool preheating. The specific method includes:

[0058] The difference between the temperature of any critical thermally sensitive point at any given time and the initial temperature is taken as the real-time temperature rise of that critical thermally sensitive point at that time; the average of the real-time temperature rises of all critical thermally sensitive points at that time is taken as the real-time temperature rise at that time.

[0059] Furthermore, the radial thermal sensitivity coefficient is obtained based on the radial displacement change of the standard test bar during the machine tool preheating process and the temperature rise change during the machine tool preheating process.

[0060] As an example, in this embodiment, during the machine tool preheating process, the radial displacement change (offset) of the standard test bar at each moment is measured in real time, as well as the temperature change relative to the temperature before machine tool processing (real-time temperature minus the temperature before processing). A linear fit is performed using the least squares method, where the horizontal axis represents the temperature change and the vertical axis represents the radial displacement change. The slope of the obtained fitted line is used as the radial thermal sensitivity coefficient.

[0061] Furthermore, the product of the real-time temperature rise at any given moment and the radial thermal sensitivity coefficient is taken as the real-time thermal drift compensation amount at that moment; wherein the real-time thermal drift compensation amount is a length quantity, with the unit being micrometers.

[0062] Preferably, in one embodiment of the present invention, the total radial compensation at each moment is obtained by combining the real-time tool deflection compensation amount, and the specific method includes:

[0063] The sum of the real-time tool deflection compensation and the real-time thermal drift compensation at any given moment on either spindle is taken as the total radial compensation of that spindle at that moment. Similarly, the total radial compensation of the other spindle at that moment is obtained, and so on, to obtain the total radial compensation of the left and right spindles at each moment. It should be noted that the real-time tool deflection compensation and the real-time thermal drift compensation of the left and right spindles need to be obtained separately, so the obtained total radial compensation also corresponds to the left and right spindles respectively.

[0064] Thus, the total radial compensation at each moment is obtained.

[0065] Step S003: Based on the sliding window, obtain the maximum cross-correlation coefficient of the left and right spindle torque signals at each moment, and then combine the cross-correlation coefficient changes during the machining process to obtain the approach intensity at each moment; obtain the peak phase lag of the cross-correlation and combine it with threshold analysis to obtain the axis skew coefficient at each moment; perform low-pass filtering on the approach intensity to obtain the local trend estimate at each moment, and then obtain the trend deviation at each moment and combine it with the axis skew coefficient to obtain the instability measurement coefficient at each moment; combine the approach intensity to obtain the approach intensity compensation weight at each moment.

[0066] It should be noted that in the docking process, due to the closed environment of deep hole machining, the sensor is difficult to install and the response speed of precise geometric measurement is slow. Directly measuring the distance and axis misalignment of the two tools cannot meet the real-time control requirements. However, the spindle torque signal has the characteristics of being easy to obtain and sensitive to changes in cutting state. Its cross-correlation characteristics can indirectly reflect the mechanical coupling relationship between the two tools. Therefore, by analyzing the cross-correlation characteristics of the dual spindle torque signals, a fuzzy state index is constructed.

[0067] Preferably, in one embodiment of the present invention, the maximum cross-correlation coefficient of the left and right spindle torque signals at each moment is obtained based on a sliding window, and then the approach intensity at each moment is obtained by combining the changes in the cross-correlation coefficient during the machining process. The specific method includes:

[0068] For any moment in either of the two torque signals of the left and right spindles, a sliding window is used to obtain a segment of local signal corresponding to that moment in each of the two torque signals of the left and right spindles. In this embodiment, the sliding window is processed in 0.1-second intervals. The moment in question is taken as the last moment in the window, and a segment of local signal corresponding to that moment in each torque signal of the left and right spindles is obtained through the sliding window. It is worth noting that if the moment in question is close to the first moment in the machining process, the local signal is constructed based on the actual moment in the process. A time delay variable and its value range are preset. In this embodiment, the value range covers a reasonable physical propagation scenario and is not specifically restricted. The cross-correlation coefficients between the two segments of local signal under different time delay variables are obtained, and the maximum value among all cross-correlation coefficients under all time delay variables is taken as the maximum value of the cross-correlation coefficient at that moment.

[0069] Furthermore, based on the mean and standard deviation of the maximum cross-correlation coefficients at each time point, a cross-correlation coefficient threshold is obtained. In this embodiment, the sum obtained by adding twice the standard deviation to the mean is used as the cross-correlation coefficient threshold. According to the time sequence, the first time of the two times when the maximum cross-correlation coefficients of two consecutive times are greater than or equal to the cross-correlation coefficient threshold is used as the background cutoff time. The mean of the maximum cross-correlation coefficients of all times from the first time in the processing to the time before the background cutoff time is used as the background correlation level. The background correlation level characterizes the strength of the inherent, systematic correlation between the dual spindle torque signals in the processing stage without significant mechanical coupling.

[0070] Furthermore, for any given moment, the difference between the maximum cross-correlation coefficient at that moment and the background correlation level is used as the additional coupling strength at that moment. This additional coupling strength represents the additional coupling strength beyond the background level caused purely by the approaching distance, after the system noise floor has been removed. The difference between 1 and the background correlation level is used as the background correlation removal level. The ratio of the additional coupling strength at that moment to the background correlation removal level is used as the approaching strength at that moment. It is worth noting that before the background cutoff time is reached during the processing, a preset initial background correlation level is used for calculation. In this embodiment, the initial background correlation level is described as 0.05.

[0071] It should be noted that the convergence strength is 0 when the maximum cross-correlation coefficient equals the background correlation level, and 1 when the maximum cross-correlation coefficient reaches its theoretical maximum of 1, i.e., when the two signals are perfectly correlated; the convergence strength ranges from [value missing]. The larger the value, the closer the coupling strength is to the theoretical limit, indicating a higher probability of physical contact.

[0072] It should be further explained that, in the head-to-head machining process, the relative spatial position of the two hole axes inside the workpiece cannot be directly observed, resulting in the axes not being collinear, i.e., skewness. When the two axes are skewed, the system structural response characteristics caused by cutting will change, resulting in a phase difference in the torque fluctuation of the left and right spindles in the low-frequency range.

[0073] Preferably, in one embodiment of the present invention, the method for obtaining the cross-correlation peak phase lag and combining it with threshold analysis to obtain the axis skew coefficient at each time step includes:

[0074] Based on the maximum cross-correlation coefficients of the left and right spindle torque signals at each moment, the maximum cross-correlation coefficients at each moment and the phase lag corresponding to the peak value of the maximum cross-correlation coefficients are obtained. A maximum allowable phase lag threshold is preset. In this embodiment, the maximum allowable phase lag threshold is described as 10ms, which is set based on empirical values. Implementers can set it according to actual conditions. The ratio of the absolute value of the phase lag at any moment to the maximum allowable phase lag threshold is multiplied by the axis skew coefficient at that moment, and the resulting product is used as the axis skew coefficient at that moment. The axis skew coefficient is used to characterize the phase anomaly degree of the system's dynamic response, and is not the geometric skew obtained by direct measurement.

[0075] It should be further explained that the machining process is divided into two key stages: independent cutting and force chain coupling. The stress state of the tools differs significantly in different stages. In the independent cutting stage, the two tools bear the cutting force independently, and the deformation is caused only by their own force, which requires full compensation. However, when the two tools form a force chain coupling through the workpiece, the cutting force will be transmitted between the two tools, and the overall rigidity of the system will be improved. If the compensation amount calculated based on the independent cutting force is still applied in full, it will lead to overcompensation and cause machining deviation.

[0076] Preferably, in one embodiment of the present invention, the approach intensity is low-pass filtered to obtain local trend estimates at each time step, and then the trend deviation at each time step is obtained and combined with the axis skew coefficient to obtain the instability metric coefficient at each time step. The specific method includes:

[0077] It should be noted that after the interaction begins, the two tools form a rigid force chain through the intermediate workpiece material. The cutting force of one tool will be transmitted to the other tool through the workpiece, changing the stress state of both tools from an independent cantilever beam to a mutually supporting coupled body. The theoretical compensation amount calculated based on the independent material removal rate no longer matches the actual stress deformation. Full compensation will lead to overcompensation, so it is necessary to reduce the compensation weight.

[0078] It should be further noted that when the coupling change process reflected by the approach strength is smooth and continuous, it indicates that the force chain is formed gradually and reliably, and the system stiffness increases gradually. At this time, the weight can decay gradually. If the coupling process fluctuates violently and repeatedly, it indicates that the interaction is unstable and may be affected by chip interference or material inhomogeneity. At this time, the increase in system stiffness is unreliable, and the weight decay should be delayed to retain more compensation based on the rigid model to stabilize the tool tip position. Moreover, because the instantaneous high-frequency fluctuations of the approach strength do not reflect the physical trend, while the low-frequency trend changes represent the real coupling process.

[0079] Specifically, using a first-order low-pass filter (with a cutoff frequency set to 0.1Hz in this embodiment), the signal composed of the approach intensities at each time moment is low-pass filtered, and the data corresponding to each time moment after filtering is used as the local trend estimate for that time moment. The absolute value of the difference between the approach intensity at any time moment and the local trend estimate at that time moment is used as the trend deviation at that time moment. The product of the ratio of the trend deviation at that time moment to the local trend estimate and the axis skew coefficient at that time moment is used as the instability measurement coefficient at that time moment. It should be noted that in the ratio calculation process of this embodiment, a minimum value is added to the denominator to avoid the denominator being 0 and making the ratio meaningless. In this embodiment, the minimum value is... The description is as follows: In this embodiment, when the local trend estimate approaches 0, the local trend estimate is set to be less than or equal to... In this case, the instability metric coefficient is set to 0.

[0080] It should be noted that the larger the instability metric coefficient, the further the actual signal deviates from the stationary trend, that is, the more unstable the process.

[0081] Preferably, in one embodiment of the present invention, the method for obtaining the approach intensity compensation weight at each time step by combining the approach intensity includes:

[0082] Using the approach intensity at any given time as the base, an exponential function is constructed by adding 1 to the instability metric coefficient at that time. The approach intensity compensation weight at that time is obtained based on the exponential function, and the approach intensity compensation weight is negatively correlated with the output value of the exponential function.

[0083] As an example, this embodiment uses the difference between 1 and the output value of the exponential function as the approach intensity compensation weight at the corresponding time.

[0084] It should be noted that when the process is stable, the instability metric coefficient approaches 0, and the sum of 1 and the instability metric coefficient approaches 1, so the compensation weight decreases approximately linearly. When the process is unstable, the instability metric coefficient approaches 0, and the sum of the exponent 1 and the instability metric coefficient is greater than 1. At this time, for the same approach strength, the compensation weight will increase. The larger the compensation weight, the more independent the system is, and the theoretical compensation amount should be almost fully executed. Conversely, the smaller the compensation weight, the more the system enters a strongly coupled, stable state, and the theoretical compensation amount should be almost no longer applied, relying on the system's own rigidity.

[0085] Thus, the approach intensity compensation weights at each time point are obtained.

[0086] Step S004: Based on the total radial compensation amount and the approach strength compensation weight, obtain the final compensation amount at each moment and output the command to the machine tool CNC system, superimpose the theoretical cutting trajectory to adjust the tool tip position; based on the adjusted cutting state, form a control cycle until the docking machining is completed.

[0087] Specifically, the product of the total radial compensation amount of any spindle at any moment and its approach strength compensation weight is taken as the final compensation amount of the spindle at that moment, and the final compensation amount of the other spindle at that moment is also obtained. A digital instruction is generated for the final compensation amount and output as the final compensation instruction in real time to the machine tool control system. This is used as the additional offset for the position of the left and right spindles and superimposed with the theoretical cutting trajectory during the machining process (which has been obtained in the CNC machine tool before the current workpiece is machined) to adjust the actual position of the tool tip.

[0088] Furthermore, after the actual position of the tool tip is adjusted, the change in the tool tip position will affect the cutting state, and will affect the torque signal, load current signal and temperature changes in real time. Real-time data acquisition is carried out again, and the analysis is performed according to the above method. The subsequent compensation amount is calculated in real time and fed back to the machine tool control system for execution, realizing the control cycle until the docking machining is completed, forming a complete closed-loop control system, realizing one-stop docking machining of high-precision stepped deep long holes.

[0089] This concludes the embodiment.

[0090] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A one-stop machining method for a head high precision stepped deep hole, characterized in that, The method includes the following steps: Record the initial temperature of several key thermally sensitive points after the machine tool is preheated; acquire the left and right spindle torque signals and the temperature of each key thermally sensitive point during the current workpiece machining process; Based on the cutting cross-sectional area and actual feed rate obtained at each moment during the machining process, the real-time material removal rate at each moment is calculated. Combined with the cutting force coefficient per unit removal rate obtained from historical machining data, the real-time cutting force at each moment is obtained, and then the real-time tool deflection compensation amount at each moment is generated. Based on the temperature change at each moment and the difference from the initial temperature, and combined with the radial thermal sensitivity coefficient measured by the machine tool preheating, the real-time thermal drift compensation amount at each moment is obtained. Combined with the real-time tool deflection compensation amount, the total radial compensation amount at each moment is obtained. The maximum cross-correlation coefficient of the left and right spindle torque signals at each moment is obtained based on the sliding window, and then the approach intensity at each moment is obtained by combining the cross-correlation coefficient changes during the machining process; the peak phase lag of the cross-correlation is obtained and combined with threshold analysis to obtain the axis skew coefficient at each moment; the approach intensity is low-pass filtered to obtain the local trend estimate at each moment, and then the trend deviation at each moment is obtained and combined with the axis skew coefficient to obtain the instability measurement coefficient at each moment; the approach intensity compensation weight at each moment is obtained by combining the approach intensity. Based on the total radial compensation and the approach strength compensation weight, the final compensation amount at each moment is obtained and the command is output to the machine tool CNC system. The theoretical cutting trajectory is superimposed to adjust the tool tip position. A control loop is formed based on the adjusted cutting state until the docking machining is completed. The method for obtaining the maximum cross-correlation coefficient of the left and right spindle torque signals at each moment based on a sliding window, and then combining the cross-correlation coefficient changes during the machining process to obtain the approach intensity at each moment, includes the following specific methods: For any moment in the two torque signals of the left and right spindles, a sliding window is used to obtain a segment of local signal corresponding to each of the two torque signals of the left and right spindles at that moment. The time delay variable and its value range are preset, and the cross-correlation coefficient between the two segments of local signal under different time delay variables is obtained. The maximum value of the cross-correlation coefficient under all time delay variables is taken as the maximum value of the cross-correlation coefficient at that moment. Based on the mean and standard deviation of the maximum cross-correlation coefficients at each time point, the cross-correlation coefficient threshold is obtained; according to the time sequence, the first time point among the two times when the maximum cross-correlation coefficients of two consecutive times are greater than or equal to the cross-correlation coefficient threshold is taken as the background cutoff time; the mean of the maximum cross-correlation coefficients of all times from the first time point in the processing to the time point before the background cutoff time is taken as the background correlation level. The difference between the maximum cross-correlation coefficient at any given time and the background correlation level is used as the additional coupling strength at that time; the difference between 1 and the background correlation level is used as the background correlation removal level; and the ratio of the additional coupling strength at that time to the background correlation removal level is used as the approach strength at that time. The specific method for obtaining the axis skewness coefficient at each time point is as follows: Based on the maximum cross-correlation coefficient of the left and right spindle torque signals at each moment, obtain the maximum cross-correlation coefficient at each moment and the phase lag corresponding to the peak value of the maximum cross-correlation coefficient; A maximum permissible phase lag threshold is preset, and the ratio of the absolute value of the phase lag at that moment to the maximum permissible phase lag threshold is used as the axis skew coefficient at that moment. The specific methods for obtaining the instability metric coefficients at each time point are as follows: The absolute value of the difference between the approach strength at that moment and the local trend estimate at that moment is taken as the trend deviation at that moment. The ratio of the trend deviation at that moment to the local trend estimate, multiplied by the axis skew coefficient at that moment, is used as the instability measure coefficient at that moment. The specific method for obtaining the approach intensity compensation weights at each time point is as follows: Using the approach intensity at that moment as the base, an exponential function is constructed by adding 1 to the instability metric coefficient at that moment. The approach intensity compensation weight at that moment is obtained based on the exponential function, and the approach intensity compensation weight is negatively correlated with the output value of the exponential function.

2. The one-stop machining method for high-precision stepped deep holes according to claim 1, characterized in that, The calculation yields the real-time material removal rate at each moment. Combined with the cutting force coefficient per unit removal rate obtained from historical processing data, the real-time cutting force at each moment is calculated. The specific method includes: The product of the cutting cross-sectional area and the actual feed rate at any given moment is taken as the real-time material removal rate at that moment. Based on the cutting force of the same workpiece material and tool type as the current workpiece under different material removal rates in historical processing data, the cutting force coefficient per unit removal rate is obtained by linear regression fitting. The product of the real-time material removal rate at that moment and the cutting force coefficient per unit removal rate is taken as the real-time cutting force at that moment.

3. The one-stop machining method for high-precision stepped deep holes according to claim 1, characterized in that, The specific method for obtaining the real-time tool deflection compensation amount at each moment is as follows: The deflection coefficient is calculated based on the tool overhang length, elastic modulus, and moment of inertia of the cross section. The product of the real-time cutting force at any given moment and the deflection coefficient is taken as the real-time tool deflection compensation amount at that moment.

4. The one-stop machining method for high-precision stepped deep holes according to claim 1, characterized in that, The specific method for obtaining the real-time thermal drift compensation amount at each moment is as follows: The difference between the temperature of any critical thermally sensitive point at any given time and the initial temperature is taken as the real-time temperature rise of that critical thermally sensitive point at that time; the average of the real-time temperature rises of all critical thermally sensitive points at that time is taken as the real-time temperature rise at that time. The radial thermal sensitivity coefficient is obtained based on the radial displacement change of the standard test bar during the machine tool preheating process and the temperature rise change during the machine tool preheating process. The product of the real-time temperature rise at that moment and the radial thermal sensitivity coefficient is used as the real-time thermal drift compensation amount at that moment.

5. The one-stop machining method for high-precision stepped deep holes according to claim 1, characterized in that, The specific method for obtaining the total radial compensation at each time point is as follows: The sum of the real-time tool deflection compensation and the real-time thermal drift compensation at any moment on any side of the spindle is taken as the total radial compensation of that side of the spindle at that moment.

6. The one-stop machining method for high-precision stepped deep holes according to claim 1, characterized in that, The specific methods for obtaining the final compensation amount at each time point are as follows: The product of the total radial compensation amount of the spindle on one side at any time and its approach strength compensation weight is taken as the final compensation amount of the spindle on that side at that time, and the final compensation amount of the spindle on the other side at that time is obtained.

Images