An intelligent tool system integrating sensing, computing and control

Abstract

The present invention discloses an intelligent tool system with integrated sensing, calculation and control, comprising an intelligent tool (1), a data acquisition card (4), a numerical control system (5) and a cloud database (6); the data acquisition card (4) collects vibration and cutting force data of the intelligent tool (1) and transmits the data to the numerical control system (5); the numerical control system (5) comprises an intelligent tool database module, a cutting force and vibration data sensing module, a tool state high-precision in-situ real-time monitoring theory and calculation module, and a tool wear high-precision timely in-situ compensation and state control module; the cloud database (6) is used to store data of the numerical control system (5). The present invention can meet the existing intelligent manufacturing needs of precision / ultra-precision machining, and can compensate for tool wear in-situ, thereby improving tool life and the surface quality of the workpiece to be machined. Therefore, the industrial application and promotion value is high.

Description

Technical Field

[0001] The present invention relates to the field of precision / ultra-precision intelligent tool measurement, monitoring and control, and in particular to an intelligent tool system integrating sensing, computing and control. Background Art

[0002] Cutting tools, the "teeth" of CNC machine tools, play an irreplaceable role in the manufacturing industry. With the rise of intelligent manufacturing, multidisciplinary cross-industry integration is gaining importance. Intelligent cutting tools, which transcend traditional cutting functions, have been proposed and rapidly developed. Embedded or integrated sensors are currently the most common form of intelligent cutting tools. Compared to machine-mounted monitoring systems, intelligent cutting tools can provide real-time monitoring of force, vibration, temperature, tool breakage, and wear, reducing monitoring costs and improving machining efficiency, tool life, and surface quality.

[0003] Most of the early research on intelligent tools was on integrated sensing and computing tools, that is, they sensed the cutting force, vibration and temperature data information during the processing, and used big data, artificial intelligence and cloud computing technologies to complete data storage, feature extraction and predictive analysis functions. However, there is little research on in-situ on-machine control based on flexible hinges and piezoelectric ceramics, making it difficult to complete in-situ drive compensation for tool wear and constant-state cutting, such as surface residual stress consistency cutting and constant force cutting, which in turn affects the quality control of the workpiece.

[0004] Due to the lack of control modules in integrated sensing and computing tools, existing intelligent tools generally have problems such as inability to actively compensate, low reliability and environmental adaptability, and cannot be widely used in the intelligent processing of difficult-to-process and special hard and brittle materials. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent tool system integrating sensing, computing and control, including an intelligent tool, a data acquisition card, a numerical control system and a cloud database;

[0006] The data acquisition card collects vibration and cutting force data of the intelligent tool and transmits them to the CNC system;

[0007] The CNC system includes an intelligent tool database module, a cutting force and vibration data perception module, a tool state high-precision in-situ real-time monitoring theory and calculation module, and a tool wear high-precision timely in-situ compensation and state control module;

[0008] The cutting force and vibration data sensing module collects the cutting force and vibration data of the intelligent tool and transmits them to the intelligent tool database module;

[0009] The intelligent tool database module is used to store the cutting force, vibration data and operating condition data of the intelligent tool;

[0010] The high-precision in-situ real-time monitoring theory and calculation module of the tool status are used to calculate the wear value of the intelligent tool;

[0011] The tool wear high-precision timely in-situ compensation and state control module is used to actively control the relative vibration between the intelligent tool and the workpiece, and compensate for the dynamic stiffness under different cutting conditions;

[0012] The cloud database is used to store data of the numerical control system.

[0013] Furthermore, the intelligent tool database module includes a cutting parameter and workpiece material query unit and an intelligent tool full working condition data storage and comparison unit;

[0014] The cutting parameter and workpiece material query unit is used to query the cutting parameters corresponding to the workpiece material;

[0015] The intelligent tool full working condition data storage and comparison unit

[0016] Furthermore, when the queried workpiece material has been used, if the intelligent tool has a response time value for the abnormal working conditions of chipping or breaking of the tool, the intelligent tool alarm light will be on for a long time, and after resetting, the workpiece material will be re-entered or selected and the corresponding cutting parameters will be given;

[0017] If there is no response time value for abnormal working conditions of chipping or breaking of the intelligent tool, the operating condition data is read; the operating condition data includes cutting force and acceleration nonlinear error, frequency response and cross interference error, intelligent tool wear online measurement error, and intelligent tool wear compensation positioning accuracy value;

[0018] When the queried workpiece material is not used, the query operation stops.

[0019] Furthermore, the intelligent tool full working condition data storage and comparison unit is used to store the cutting force and vibration data of the intelligent tool during the machining process;

[0020] When the high-precision timely in-situ compensation and state control module for tool wear is completed, the intelligent tool full-working condition data storage and comparison unit also reads the cutting force and acceleration nonlinear errors, frequency response and cross-interference errors, intelligent tool wear online measurement errors and intelligent tool wear compensation positioning accuracy values, and compares them with historical records, retaining smaller error values ​​and larger intelligent tool wear compensation positioning accuracy values.

[0021] Furthermore, the cutting force and vibration data perception module includes a cutting process force and vibration signal analysis unit and a cutting force and acceleration sensor state monitoring unit;

[0022] The cutting process force and vibration signal analysis unit is used to select the arrangement position of the force sensor and the acceleration sensor;

[0023] The force sensor and acceleration sensor are both arranged at the tool bar and tool handle of the intelligent tool, provided that the force and acceleration signals meet the nonlinear error of less than or equal to 0.3% FS and the frequency response is greater than or equal to 30kHz; FS represents full scale;

[0024] The cutting force and acceleration sensor condition monitoring unit stores a multi-dimensional force and acceleration cross-coupling analytical model;

[0025] The cutting force and acceleration sensor state monitoring unit inputs the data monitored by the force sensor and the acceleration sensor into a multi-dimensional force and acceleration cross-coupling analytical model, suppresses sensor cross interference, and processes to obtain cutting force and vibration data.

[0026] The steps to establish the multi-dimensional force and acceleration cross-coupling analytical model include:

[0027] 1) The design coordinates of the acceleration sensor coincide with the setting coordinates of the intelligent tool;

[0028] 2) The design coordinates of the force sensor coincide with the design coordinates of the acceleration sensor;

[0029] 3) With the help of acceleration information, the coupling of inertial force to the force sensor is separated, and a multi-dimensional force and acceleration cross-coupling analytical model is established through iterative decoupling.

[0030] Furthermore, the high-precision in-situ real-time monitoring theory and calculation module of the tool status includes a tool dynamic evolution multi-source sensor information mapping unit, a tool breakage monitoring unit for abnormal feature recognition, and a multi-condition tool wear prediction unit for transfer learning;

[0031] The tool dynamic evolution multi-source sensor information mapping unit is used to monitor the tool multi-dimensional force and acceleration data characteristic data;

[0032] The tool breakage monitoring unit for abnormal feature recognition builds a multi-blade tool state change cloud map based on multi-dimensional force and acceleration data, and identifies the breakage and chipping of each tooth of the tool. If the response time of the intelligent tool chipping and tool breakage abnormal working condition is less than or equal to t, it enters the multi-condition tool wear prediction unit of transfer learning. Otherwise, it continues to build the multi-blade tool state change cloud map;

[0033] The multi-working condition tool wear prediction unit of the transfer learning stores a transfer learning model for tool wear;

[0034] The transfer learning model for tool wear is used to predict intelligent tool wear values.

[0035] Furthermore, the transfer learning model for tool wear is trained by a wear value smoothing method based on wear curve fitting.

[0036] The extracted feature point x on the wear curve i obeys the normal distribution with a mean of 0 and a variance of The probability density function is

[0037] Probability density function As shown below:

[0038]

[0039] Where X is a random vector in N-dimensional space, f(h) is the principal curve of X, ||Xf(h)|| is the Euclidean distance between X and the mapping point on the principal curve, l i is the curve length, is the variance of X from the principal curve. i is the standard deviation of X from the principal curve.

[0040] Furthermore, the tool wear high-precision timely in-situ compensation and state control module includes a tool wear high-precision in-situ compensation control unit with micro-displacement feedback, a piezoelectric driven cutting state control unit with force / torque feedback, and a piezoelectric driven cutting state control unit with vibration feedback;

[0041] The high-precision in-situ compensation control unit for tool wear with micro-displacement feedback utilizes piezoelectric hysteresis nonlinear feedback linearization and an amplitude-frequency dual-correlation equivalent disturbance compensation control strategy based on a disturbance observer to achieve tool wear compensation control.

[0042] The steps of implementing tool wear compensation control by the micro-displacement feedback high-precision in-situ tool wear compensation control include:

[0043] 1) Establish a Hammerstein dynamic model of the intelligent tool system to describe the static hysteresis nonlinearity of the intelligent tool system and the dynamic linear characteristics of the tool;

[0044] The Hammerstein dynamics model of the intelligent tool system is shown below:

[0045]

[0046] In the formula, the theoretical value Theoretical value Theoretical value Theoretical value Theoretical value Estimated value b1 = 2ζω n ; Estimated value m1, a0, a i are estimated values; Δb1, Δb2, Δm1, Δa0, Δa i is the uncertain parameter value; q1(t) is the external disturbance; ξ and ω nis the damping ratio and natural frequency; F ri [u](t) is the output value of the play operator; is the output of the intermediate hysteresis; y(t) is the output of the intelligent tool system; u(t) is the input;

[0047] 2) The nonlinear Hammerstein dynamics model is transformed into a linear system using the feedback linearization algorithm, namely:

[0048]

[0049] Where, is the uncertainty disturbance in the MPI model;

[0050] 3) Construct the relationship between output y(t) and control w(t), namely:

[0051]

[0052] Where q(t)=-Δb1y(t)-Δb2y(t)+Δb2w(t)+Δb2q2(t)+q1(t) is the equivalent perturbation;

[0053] 4) Design a three-degree-of-freedom controller, including a feedforward controller, a PID feedback controller, and an equivalent disturbance compensation controller;

[0054] The control input w(t) of the three-degree-of-freedom controller is as follows:

[0055] w(t)=w FF (t)+w FB (t)+w DC (t) (5)

[0056] Where w FF (t) is the feedforward controller input, w FB (t) is the input of PID feedback controller, w DC (t) is the input of the equivalent disturbance compensation controller;

[0057] The piezoelectric driven cutting state control unit with force / torque feedback establishes a mapping relationship between cutting force and cutting parameters, and constructs a master-slave contact force feedback control strategy;

[0058] The vibration feedback piezoelectric driven cutting state control unit stores a vibration prediction model and a chatter prediction analysis model;

[0059] The vibration prediction model and chatter prediction analysis model are used to predict the vibration and chatter of the intelligent tool respectively;

[0060] The vibration feedback piezoelectric driven cutting state control unit actively controls the relative vibration between the intelligent tool and the workpiece according to vibration and chatter, and compensates for the dynamic stiffness under different cutting conditions.

[0061] Further, a charge amplifier is included;

[0062] The charge amplifier amplifies the vibration and cutting force data of the intelligent tool and transmits the data to the data acquisition card.

[0063] Further, it also includes a signal line;

[0064] The intelligent tool and the charge amplifier, the data acquisition card and the numerical control system, and the charge amplifier and the data acquisition card transmit data via signal lines.

[0065] The technical effects of the present invention are undoubted, and the beneficial effects of the present invention are as follows:

[0066] 1) As traditional cutting tools are difficult to adapt to the current intelligent manufacturing environment, an intelligent tool system with integrated sensing, computing and control is constructed from four aspects: intelligent tool database, cutting force and vibration data perception, high-precision in-situ real-time monitoring theory and calculation of tool status, and high-precision timely in-situ compensation and status control of tool wear. The system involves: intelligent query of cutting process, storage and comparison of data under all working conditions, cutting force and vibration signal analysis, processing status monitoring, multi-source sensor information mapping, abnormal status identification, tool wear prediction, micro-displacement feedback in-situ compensation, force / torque feedback piezoelectric drive control, and vibration feedback piezoelectric drive control.

[0067] 2) Through the analysis of the intelligent tool database module, an innovative intelligent tool with integrated sensing, computing and control is proposed and implemented. Based on the principle that low-precision data obeys high-precision data, historical records retain the optimal precision data, providing data support for the continuous optimization of subsequent data.

[0068] 3) Through the analysis of the cutting force and vibration data perception module, without changing the rigidity of the tool, a miniature force sensor is innovatively introduced to measure the cutting force, and the nonlinear error and cross-interference suppression of the force and vibration sensor signals are taken into consideration. The nonlinear error of the force and acceleration signals must be ≤0.3% FS, the frequency response must be ≥30kHz, and the cross-interference error must be ≤2.0%.

[0069] 4) Through the analysis of the theory and calculation modules of high-precision in-situ real-time monitoring of tool status, a high-throughput computing system is constructed, multi-source data features are extracted, and threshold segmentation, support vector machine, Gaussian clustering and cross-correlation algorithms are used to reveal the spatiotemporal evolution mechanism of tool wear and the dynamic evolution multi-source sensor information mapping mechanism; a dynamic threshold anomaly detection method is proposed to improve the recognition ability and response speed of abnormal features of breakage and chipping, and the response time of abnormal working conditions of intelligent tool chipping and broken tool is ≤5ms; a transfer learning model for tool wear is constructed to ensure the prediction accuracy of tool wear value, and the online measurement error of intelligent tool wear is ≤2μm.

[0070] 5) Through the analysis of the high-precision timely in-situ compensation and state control module of tool wear, a piezoelectric hysteresis nonlinear feedback linearization and an amplitude-frequency dual-correlation equivalent disturbance compensation control strategy based on the disturbance observer are proposed to realize tool wear compensation control, and the positioning accuracy of intelligent tool wear compensation is ≤0.2μm.

[0071] 6) This intelligent tool system innovatively uses micro force sensors to monitor the machining process in real time. Compared with strain and piezoelectric sensors, micro force sensors are easier to integrate and significantly reduce production costs under the condition that the rigidity of the tool / tool ​​holder remains unchanged.

[0072] 7) This intelligent tool system involves multi-sensor fusion of force and acceleration sensors, which can fully reflect the instantaneous state of the machining process through information complementarity to improve the system monitoring accuracy.

[0073] 8) This intelligent tool system proposes an innovative concept of integrated sensing, computing and control, which can achieve accurate prediction of tool wear, focus on data-driven intelligent control, and improve high-precision and timely in-situ compensation.

[0074] 9) This intelligent tool system can meet the existing intelligent manufacturing needs of precision / ultra-precision machining, and can compensate for tool wear in situ, thereby improving tool life and the surface quality of the workpiece to be machined. Therefore, it has high value for industrial application promotion.

[0075] 10) The present invention proposes a sensor-computing-controlled step-by-step modular progressive control strategy to accurately control the single-step progressive accuracy index and provide support for improving the accuracy of the intelligent tool system under all working conditions.

[0076] This intelligent tool system can meet the requirements of workpiece surface quality and tool long life management during precision / ultra-precision single machining, which improves machining efficiency and reduces machining energy consumption. Therefore, it has broad application prospects in the field of high-performance intelligent manufacturing. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 FIG1 is a hardware configuration diagram of the intelligent tool system of the present invention;

[0078] Figure 2This is the overall architecture diagram of the intelligent tool system of the present invention;

[0079] Figure 3 This is a diagram showing the relationship between modules of the intelligent tool database of the present invention;

[0080] Figure 4 This is a diagram showing the relationship between the cutting force and vibration data sensing module of the present invention;

[0081] Figure 5 This is a relationship diagram between the theory and calculation module of high-precision in-situ real-time monitoring of tool status of the present invention;

[0082] Figure 6 This is a relationship diagram of the tool wear high-precision timely in-situ compensation and state control module of the present invention;

[0083] Figure 7 is the Hammerstein dynamics model;

[0084] Figure 8 is the linear relationship between input ω(t) and output y(t);

[0085] Figure 9 The relationship between the various parts of the three-degree-of-freedom controller;

[0086] In the figure: 1. Intelligent tool; 2. Charge amplifier; 3. Signal line; 4. Data acquisition card; 5. CNC system; 6. Cloud database. DETAILED DESCRIPTION

[0087] The present invention will be further described below with reference to the following examples, but it should not be understood that the scope of the present invention is limited to the following examples. Without departing from the above technical ideas of the present invention, various substitutions and modifications can be made according to common technical knowledge and customary means in the art, and all should be included in the scope of protection of the present invention.

[0088] Example 1:

[0089] See also Figures 1-9 , a sensing, computing and controlling integrated intelligent tool system, comprising an intelligent tool 1, a data acquisition card 4, a numerical control system 5 and a cloud database 6;

[0090] The data acquisition card 4 collects the vibration and cutting force data of the intelligent tool 1 and transmits it to the numerical control system 5;

[0091] The numerical control system 5 includes an intelligent tool database module, a cutting force and vibration data perception module, a tool state high-precision in-situ real-time monitoring theory and calculation module, and a tool wear high-precision timely in-situ compensation and state control module;

[0092] The cutting force and vibration data sensing module collects the cutting force and vibration data of the intelligent tool 1 and transmits it to the intelligent tool database module;

[0093] The intelligent tool database module is used to store the cutting force, vibration data and operating condition data of the intelligent tool 1;

[0094] The high-precision in-situ real-time monitoring theory and calculation module of the tool status are used to calculate the wear value of the intelligent tool 1;

[0095] The tool wear high-precision timely in-situ compensation and state control module is used to actively control the relative vibration between the intelligent tool and the workpiece, and compensate for the dynamic stiffness under different cutting conditions;

[0096] The cloud database 6 is used to store data of the numerical control system 5 .

[0097] Example 2:

[0098] A sensing, computing and controlling integrated intelligent tool system, the technical content of which is the same as that of Example 1, further, the intelligent tool database module includes a cutting parameter and workpiece material query unit, and an intelligent tool full working condition data storage and comparison unit;

[0099] The cutting parameter and workpiece material query unit is used to query the cutting parameters corresponding to the workpiece material;

[0100] The intelligent tool full working condition data storage and comparison unit

[0101] Example 3:

[0102] A sensing, computing and controlling integrated intelligent tool system, having the same technical content as any one of embodiments 1-2, further comprising: when the queried workpiece material has been used, if the intelligent tool 1 has an abnormal working condition of chipping or breaking and has a response time value, the intelligent tool 1 alarm light is permanently on, and after resetting, the workpiece material is re-entered or selected and the corresponding cutting parameters are given;

[0103] If there is no response time value for the abnormal working condition of the chipping or breaking of the intelligent tool 1, the operating condition data is read; the operating condition data includes the nonlinear error of cutting force and acceleration, frequency response and cross interference error, online measurement error of the wear of the intelligent tool 1, and the positioning accuracy value of the intelligent tool wear compensation;

[0104] When the queried workpiece material is not used, the query operation stops.

[0105] Example 4:

[0106] A sensing, computing and controlling integrated intelligent tool system, having the same technical content as any one of embodiments 1-3, wherein the intelligent tool full-working-condition data storage and comparison unit is used to store cutting force and vibration data of the intelligent tool 1 during machining;

[0107] When the high-precision timely in-situ compensation and state control module for tool wear is completed, the intelligent tool full-working condition data storage and comparison unit also reads the cutting force and acceleration nonlinear errors, frequency response and cross-interference errors, intelligent tool wear online measurement errors and intelligent tool wear compensation positioning accuracy values, and compares them with historical records, retaining smaller error values ​​and larger intelligent tool wear compensation positioning accuracy values.

[0108] Example 5:

[0109] A sensing, computing and controlling integrated intelligent tool system, having the same technical content as any one of embodiments 1-4, further comprising: a cutting force and vibration data sensing module comprising a cutting process force and vibration signal analysis unit and a cutting force and acceleration sensor state monitoring unit;

[0110] The cutting process force and vibration signal analysis unit is used to select the arrangement position of the force sensor and the acceleration sensor;

[0111] The force sensor and acceleration sensor are both arranged at the tool bar and tool handle of the intelligent tool 1, provided that the force and acceleration signals meet the nonlinear error of less than or equal to 0.3% FS and the frequency response is greater than or equal to 30kHz; FS represents full scale;

[0112] The cutting force and acceleration sensor condition monitoring unit stores a multi-dimensional force and acceleration cross-coupling analytical model;

[0113] The cutting force and acceleration sensor state monitoring unit inputs the data monitored by the force sensor and the acceleration sensor into a multi-dimensional force and acceleration cross-coupling analytical model, suppresses sensor cross interference, and processes to obtain cutting force and vibration data.

[0114] Example 6:

[0115] A sensing, computing and controlling integrated intelligent tool system, the technical content of which is the same as any one of embodiments 1-5, further comprising the steps of establishing a multi-dimensional force and acceleration cross-coupling analytical model:

[0116] 1) The design coordinates of the acceleration sensor coincide with the setting coordinates of the intelligent tool;

[0117] 2) The design coordinates of the force sensor coincide with the design coordinates of the acceleration sensor;

[0118] 3) With the help of acceleration information, the coupling of inertial force to the force sensor is separated, and a multi-dimensional force and acceleration cross-coupling analytical model is established through iterative decoupling.

[0119] Example 7:

[0120] A sensing, computing, and controlling integrated intelligent tool system, having the same technical content as any one of Examples 1-6, further comprising a high-precision in-situ real-time monitoring theory and calculation module for tool status, including a tool dynamic evolution multi-source sensing information mapping unit, an abnormal feature recognition tool damage monitoring unit, and a transfer learning multi-condition tool wear prediction unit;

[0121] The tool dynamic evolution multi-source sensor information mapping unit is used to monitor the tool multi-dimensional force and acceleration data characteristic data;

[0122] The tool breakage monitoring unit for abnormal feature recognition builds a multi-blade tool state change cloud map based on multi-dimensional force and acceleration data, and identifies the breakage and chipping of each tooth of the tool. If the response time of the intelligent tool chipping and tool breakage abnormal working condition is less than or equal to t, it enters the multi-condition tool wear prediction unit of transfer learning. Otherwise, it continues to build the multi-blade tool state change cloud map;

[0123] The multi-working condition tool wear prediction unit of the transfer learning stores a transfer learning model for tool wear;

[0124] The transfer learning model for tool wear is used to predict intelligent tool wear values.

[0125] Example 8:

[0126] A sensing, computing and controlling integrated intelligent tool system, the technical content of which is the same as any one of Examples 1-7, further, the transfer learning model for tool wear is trained by a wear value smoothing method based on wear curve fitting.

[0127] Example 9:

[0128] A sensing, computing and controlling integrated intelligent tool system, the technical content is the same as any one of embodiments 1-8, further, the extracted feature point x on the wear curve i obeys the normal distribution, its mean is 0, and the variance is The probability density function is

[0129] Probability density function As shown below:

[0130]

[0131] Where X is a random vector in N-dimensional space, f(h) is the principal curve of X, ||Xf(h)|| is the Euclidean distance between X and the mapping point on the principal curve, l i is the curve length, is the variance of X from the principal curve. i is the standard deviation of X from the principal curve.

[0132] Example 10:

[0133] A sensing, computing and control integrated intelligent tool system, having the same technical content as any one of Examples 1-9, further comprising a high-precision timely in-situ compensation and state control module for tool wear, comprising a high-precision in-situ compensation control unit for tool wear with micro-displacement feedback, a piezoelectrically driven cutting state control unit with force / torque feedback, and a piezoelectrically driven cutting state control unit with vibration feedback;

[0134] The high-precision in-situ compensation control unit for tool wear with micro-displacement feedback utilizes piezoelectric hysteresis nonlinear feedback linearization and an amplitude-frequency dual-correlation equivalent disturbance compensation control strategy based on a disturbance observer to achieve tool wear compensation control.

[0135] The piezoelectric driven cutting state control unit with force / torque feedback establishes a mapping relationship between cutting force and cutting parameters, and constructs a master-slave contact force feedback control strategy;

[0136] The vibration feedback piezoelectric driven cutting state control unit stores a vibration prediction model and a chatter prediction analysis model;

[0137] The vibration prediction model and chatter prediction analysis model are used to predict the vibration and chatter of the intelligent tool 1 respectively;

[0138] The vibration feedback piezoelectric driven cutting state control unit actively controls the relative vibration between the intelligent tool and the workpiece according to vibration and chatter, and compensates for the dynamic stiffness under different cutting conditions.

[0139] Example 11:

[0140] A sensing, computing and controlling integrated intelligent tool system, the technical content of which is the same as any one of embodiments 1-10, further, the micro-displacement feedback tool wear high-precision in-situ compensation control unit implements tool wear compensation control steps including:

[0141] 1) High-precision in-situ compensation control unit for tool wear with micro-displacement feedback:

[0142] ① First, the composite dynamic characteristics of piezoelectric ceramics and cutting tools are analyzed, and a nonlinear-linear cascade Hammerstein dynamic model of the intelligent cutting tool mechanism is established. This mathematical model can describe the static hysteresis nonlinearity of the intelligent cutting tool system and the dynamic linear characteristics of the cutting tool.

[0143] Use the MPI model to describe the static hysteresis nonlinearity of piezoelectric ceramics:

[0144]

[0145] The dynamic linear characteristics of the tool are described by a second-order linear system model:

[0146]

[0147] Hammerstein dynamics model:

[0148]

[0149] ② Nonlinear model design controller is relatively complex, so it is convenient to linearize the nonlinear model and use feedback nonlinearization algorithm to convert the nonlinear system into a linear system. The linear system should have a simple linear relationship between the new input ω(t) and the output y(t):

[0150] Therefore, it is easy to design a suitable control input ω(t) to control the tracking characteristics of the output y(t).

[0151] From formula (4), the nonlinear relationship between y(t) and u(t) can be deduced as

[0152]

[0153] If the signal u(t) is designed by the feedback linearization algorithm, its expression is:

[0154]

[0155] Where ω(t) is the new input of the feedback linearization algorithm, and f(t) is the nonlinear part of (8), which is expressed as

[0156]

[0157] Substituting u(t) from (6) into (5), we get the relationship between y(t) and ω(t):

[0158]

[0159] Clearly, there is a simple linear relationship between the output y(t) and the new input ω(t). Therefore, the static hysteresis nonlinearity of the intelligent tool system is eliminated.

[0160] ③The nonlinear system is transformed into a linear system through feedback linearization. Considering the uncertainty of the model and external disturbances, the Hammerstein model of the intelligent tool system can be described as

[0161]

[0162] in,

[0163] In the formula, the theoretical value Theoretical value Theoretical value Theoretical value Theoretical value Estimated value b1 = 2ζω n ; Estimated value m1, a0, a i are estimated values; Δb1, Δb2, Δm1, Δa0, Δa i is the uncertain parameter value; q1(t) is the external disturbance; ξ and ω n is the damping ratio and natural frequency; F ri [u](t) is the output value of the play operator; is the output of the intermediate hysteresis; y(t) is the output of the intelligent tool system; u(t) is the input;

[0164] The nonlinear system is transformed into a linear system using the feedback linearization algorithm. Equation (9) can be described as

[0165]

[0166] in is the uncertainty disturbance in the MPI model.

[0167] Therefore, the relationship between y(t) and w(t) can be obtained from formula (10):

[0168]

[0169] in,

[0170] q(t)=-Δb1y(t)-Δb2y(t)+Δb2w(t)+Δb2q2(t)+q1(t), which is an equivalent perturbation.

[0171] After performing equivalent disturbance analysis, a three-degree-of-freedom controller is designed. The controller consists of three parts: feedforward control, PID feedback control, and equivalent disturbance compensation control. The relationship is as follows:

[0172] The control input ω(t) has the form

[0173] ω(t)=ω FF (t)+ω FB (t)+ω DC (t) (12)

[0174] where ω FF (t) is the feedforward control, ω FB (t) is PID feedback control, ω DC (t) is the equivalent disturbance compensation control.

[0175] Feedforward controller design: Ignore the equivalent disturbance q(t) and use the reference signal and its derivative to design the feedforward control input ω in a model without considering the equivalent disturbance. FF (t).

[0176] Feedback Controller Design: Determining the Feedforward Control Input ω FF (t), without considering the error system of equivalent disturbance, the pole configuration method is used to design the PID feedback control input ω FB (t)

[0177] Equivalent disturbance compensation control design: Establish the optimal equivalent disturbance compensation. To eliminate high-frequency uncertainty and noise, design a disturbance compensator, which is implemented by two first-order low-pass filters in series. Calculate the equivalent disturbance in the s domain to obtain the input ω of the equivalent disturbance compensation control unit. DC (t).

[0178] Example 12:

[0179] A sensing, computing and controlling integrated intelligent tool system, having the same technical content as any one of embodiments 1-11, further comprising a charge amplifier 2;

[0180] The charge amplifier 2 amplifies the vibration and cutting force data of the intelligent tool 1 and transmits the amplified data to the data acquisition card 4 .

[0181] Example 13:

[0182] A sensing, computing and controlling integrated intelligent tool system, having the same technical content as any one of embodiments 1-12, further comprising a signal line 3;

[0183] The intelligent tool 1 and the charge amplifier 2 , the data acquisition card 4 and the numerical control system 5 , and the charge amplifier 2 and the data acquisition card 4 transmit data via the signal line 3 .

[0184] Example 14:

[0185] A sensing, computing and controlling integrated intelligent tool system, the hardware structure of which includes an intelligent tool 1, a charge amplifier 2, a signal line 3, a data acquisition card 4, a numerical control system 5 and a cloud database 6;

[0186] For details, see Figure 2 The overall architecture includes an intelligent tool database module, a cutting force and vibration data perception module, a high-precision in-situ real-time monitoring theory and calculation module for tool status, and a high-precision timely in-situ compensation and status control module for tool wear;

[0187] See also Figure 3 , the intelligent tool database module includes a cutting parameter and workpiece material query unit and an intelligent tool full working condition data storage and comparison unit;

[0188] The cutting parameter and workpiece material query unit includes inputting or selecting the workpiece material and giving the cutting parameters in the human-machine interface. When the control center inquires about the use of the corresponding workpiece material and cutting parameters, if the intelligent tool has a response time value for abnormal working conditions such as chipping and tool breakage, the red alarm light of the intelligent tool is on for a long time, and after resetting, it returns to the human-machine interface to input or select the workpiece material and give the cutting parameters; if the intelligent tool has no response time value for abnormal working conditions such as chipping and tool breakage, the history bar reads the control center's cutting force and acceleration nonlinear error, frequency response and cross-interference error, intelligent tool wear online measurement error and intelligent tool wear compensation positioning accuracy value; when the control center inquires about the use of the corresponding workpiece material and cutting parameters, the query operation stops;

[0189] The intelligent tool full working condition data storage and comparison unit includes the following: the intelligent tool measurement mode is turned on, the cutting force and vibration data are collected and stored in real time during the processing, and when the cutting force and vibration data perception module is completed, it enters the tool state high-precision in-situ real-time monitoring theory and calculation module; when the cutting force and vibration data perception module is not completed, it is re-run under the current module; when the tool state high-precision in-situ real-time monitoring theory and calculation module is completed, it enters the tool wear high-precision timely in-situ compensation and state control module; when the tool state high-precision in-situ real-time monitoring theory and calculation module is not running When it is completed, re-run under the current module; when the tool wear high-precision timely in-situ compensation and state control module is completed, a new record column is created to read the control center's cutting force and acceleration nonlinear errors, frequency response and cross-interference errors, intelligent tool wear online measurement errors and intelligent tool wear compensation positioning accuracy values, the new record column read values ​​are compared with the historical record column read values, the best value is selected as the subsequent comparison data, and the eliminated values ​​are saved in the control center's current cutting parameters and stacked layer by layer; if the tool wear high-precision timely in-situ compensation and state control module has not been completed, re-run under the current module;

[0190] See also Figure 4 , the cutting force and vibration data perception module includes a cutting process force and vibration signal analysis unit and a cutting force and acceleration sensor state monitoring unit;

[0191] The cutting process force and vibration signal analysis unit includes analyzing the force and vibration process of the intelligent tool, resolving the tool tip force and vibration based on kinematic analysis theory, conducting numerical simulation research on the tool tip, tool rod and tool handle processing process, obtaining the position of the sensor applied to the tool rod and tool handle based on deformation data and cloud map, optimizing the force and acceleration sensor without changing the rigidity of the tool handle and tool rod, installing a miniature force sensor at the force sensitive position of the tool rod, and installing an acceleration sensor at the vibration sensitive position of the tool handle. When the force and acceleration signal meet the nonlinear error less than or equal to 0.3% FS and the frequency response greater than or equal to 30kHz, it enters the cutting force and acceleration sensor state monitoring unit; when the force and acceleration signal do not meet the nonlinear error less than or equal to 0.3% FS and the frequency response greater than or equal to 30kHz, the force and acceleration sensor and its installation position are re-optimized;

[0192] The cutting force and acceleration sensor state monitoring unit includes force and vibration signals belonging to multi-signal fusion, which has a high degree of synergy and complementarity. A multi-dimensional force and acceleration cross-coupling analytical model is established to achieve cross-interference suppression of the embedded integrated sensor of the intelligent tool, and a high-precision decoupling algorithm is developed to suppress the cross-interference of the multi-dimensional force and acceleration sensors. Based on the efficient model and algorithm, the multi-signal fusion state monitoring is analyzed. When the cross-interference error of the force and acceleration signals is less than or equal to 2.0%, the signal accuracy and stability of the intelligent tool cutting process are improved; when the cross-interference error of the force and acceleration signals is greater than 2.0%, the multi-dimensional force and acceleration cross-coupling analytical model is re-established, a high-precision decoupling algorithm is developed, and multi-signal fusion state monitoring analysis is carried out;

[0193] See also Figure 5 The high-precision in-situ real-time monitoring theory and calculation module of the tool status includes a tool dynamic evolution multi-source sensor information mapping unit, a tool breakage monitoring unit based on abnormal feature recognition, and a multi-condition tool wear prediction unit based on transfer learning;

[0194] The tool dynamic evolution multi-source sensor information mapping unit includes building a high-throughput computing system to achieve accurate multi-dimensional information collection, studying a multi-source sensor information preprocessing method, using wavelet transform to improve the signal-to-noise ratio of multi-source information, and realizing multi-source data feature extraction based on empirical mode decomposition. Through threshold segmentation and support vector machine algorithm, a state feature library for accurate characterization of tool wear, chipping and tool breakage dynamic evolution is established. Gaussian clustering and cross-correlation methods are used to build a multi-source information feature fusion strategy based on sensitive factors to reveal the tool dynamic evolution multi-source sensor information mapping mechanism.

[0195] The tool breakage monitoring unit for abnormal feature recognition includes adopting a time-frequency domain perturbation and random slicing method to improve sample complexity, optimize the distribution and proportion of broken and chipped blade states, reduce the imbalance of chipping blade data, propose a dynamic threshold anomaly detection method, and perform normal / abnormal state classification to improve the recognition ability and response speed of broken and chipped blade abnormal features, construct a multi-blade tool state change cloud map based on multi-dimensional force and acceleration data, and realize high-speed intelligent recognition of each tooth breakage and chipping by monitoring the change of characteristic values. When the response time of the abnormal working condition of chipping and broken blade of intelligent tool is less than or equal to 5ms, the multi-condition tool wear prediction unit of transfer learning is entered; when the response time of the abnormal working condition of chipping and broken blade of intelligent tool is greater than 5ms, the time-frequency domain perturbation and random slicing and dynamic threshold anomaly detection method are restarted and a multi-blade tool state change cloud map is constructed;

[0196] The multi-condition tool wear prediction unit of the transfer learning includes constructing a transfer learning model for tool wear, studying the influence of hyperparameters on model performance, suppressing overfitting, exploring the adaptive feature learning performance of the deep learning network, ensuring the prediction accuracy of the tool wear value of the fully connected network, establishing an adaptive layer in the prediction network, narrowing the edge distribution difference between the source domain and the target domain, improving the tool wear prediction accuracy through model iteration and dynamic adjustment, and proposing a wear value smoothing method based on wear curve fitting to achieve effective suppression of prediction deviation in view of the problem of jitter in the wear value predicted by the model. When the online measurement error of the intelligent tool wear is less than or equal to 2μm, the accuracy of the intelligent tool cutting process signal is improved; when the online measurement error of the intelligent tool wear is greater than 2μm, the transfer learning model for tool wear is reconstructed, an adaptive layer is established in the prediction network, and a wear value smoothing method based on wear curve fitting is proposed to achieve effective suppression of prediction deviation;

[0197] See also Figure 6 The tool wear high-precision timely in-situ compensation and state control module includes a tool wear high-precision in-situ compensation control unit with micro-displacement feedback, a piezoelectric driven cutting state control unit with force / torque feedback, and a piezoelectric driven cutting state control unit with vibration feedback;

[0198] The high-precision in-situ compensation control unit for tool wear with micro-displacement feedback includes a displacement compensation method based on a micro-displacement sensor and a piezoelectric drive, which reveals the piezoelectric drive creep, hysteresis nonlinearity and mechanical vibration dynamic characteristics of the intelligent tool, studies the optimal problem of nonlinear model parameters, realizes error minimization parameter identification based on a nonlinear parameter identification algorithm, proposes a piezoelectric hysteresis nonlinear feedback linearization and an amplitude-frequency dual-correlation equivalent disturbance compensation control strategy based on a disturbance observer, realizes tool wear compensation control, and carries out piezoelectric drive cutting state regulation with force / torque / vibration feedback when the positioning accuracy of the intelligent tool wear compensation is less than or equal to 0.2 mm; when the positioning accuracy of the intelligent tool wear compensation is greater than 0.2 mm, studies the nonlinear model parameter optimization and proposes a piezoelectric hysteresis nonlinear feedback linearization and an amplitude-frequency dual-correlation equivalent disturbance compensation control strategy based on a disturbance observer;

[0199] The force / torque feedback piezoelectric driven cutting state control unit includes research on intelligent tool control technology based on in-situ micro force sensors and piezoelectric drive, establishing a mapping relationship between cutting force and cutting parameters, proposing a master-slave contact force feedback control strategy with high stiffness and robustness based on complex working condition disturbances such as machining deformation, surface damage and process parameter optimization, and researching cutting state control and process optimization methods based on force / torque feedback control to improve the machining stability of intelligent tool cutting accuracy and surface quality;

[0200] The vibration feedback piezoelectric driven cutting state control unit includes studying active vibration suppression control technology based on acceleration sensors and piezoelectric drives, establishing vibration prediction models and chatter prediction analysis models, and proposing real-time active suppression control strategies of robust interference suppression and robust mixed sensitivity, so as to realize active control of relative vibration between intelligent tools and workpieces during machining and dynamic stiffness compensation under different cutting conditions, effectively improving machining accuracy, efficiency and workpiece quality, extending tool life, and improving the precision control level of intelligent tools.

Claims

1. A sensing, computing and control integrated intelligent tool system, characterized in that: It includes an intelligent tool (1), a data acquisition card (4), a numerical control system (5) and a cloud database (6); The data acquisition card (4) collects vibration and cutting force data of the intelligent tool (1) and transmits the data to the numerical control system (5); The numerical control system (5) includes an intelligent tool database module, a cutting force and vibration data perception module, a tool state high-precision in-situ real-time monitoring theory and calculation module, and a tool wear high-precision timely in-situ compensation and state control module; The cutting force and vibration data sensing module collects the cutting force and vibration data of the intelligent tool (1) and transmits the data to the intelligent tool database module; The intelligent tool database module is used to store cutting force, vibration data and operating condition data of the intelligent tool (1); The high-precision in-situ real-time monitoring theory and calculation module of the tool state are used to calculate the wear value of the intelligent tool (1); The tool wear high-precision timely in-situ compensation and state control module is used to actively control the relative vibration between the intelligent tool and the workpiece, and compensate for the dynamic stiffness under different cutting conditions; The cloud database (6) is used to store data of the numerical control system (5); The tool wear high-precision timely in-situ compensation and state control module includes a tool wear high-precision in-situ compensation control unit with micro-displacement feedback, a piezoelectric driven cutting state control unit with force / torque feedback, and a piezoelectric driven cutting state control unit with vibration feedback; The high-precision in-situ compensation control unit for tool wear with micro-displacement feedback utilizes piezoelectric hysteresis nonlinear feedback linearization and an amplitude-frequency dual-correlation equivalent disturbance compensation control strategy based on a disturbance observer to achieve tool wear compensation control. The steps of implementing tool wear compensation control by the micro-displacement feedback high-precision in-situ tool wear compensation control include: 1) Establish a Hammerstein dynamic model of the intelligent tool system to describe the static hysteresis nonlinearity of the intelligent tool system and the dynamic linear characteristics of the tool; The Hammerstein dynamics model of the intelligent tool system is shown below: In the formula, the theoretical value Theoretical value Theoretical value Theoretical value Theoretical value Estimated value b1 = 2ζω n ; Estimated value m1, a0, a i are estimated values; Δb1, Δb2, Δm1, Δa0, Δa i is the uncertain parameter value; q1(t) is the external disturbance; ξ and ω n is the damping ratio and natural frequency; F ri [u](t) is the output value of the play operator; is the output of the intermediate hysteresis; y(t) is the output of the intelligent tool system; u(t) is the input; 2) The nonlinear Hammerstein dynamics model is transformed into a linear system using the feedback linearization algorithm, namely: Where, is the uncertainty disturbance in the MPI model; 3) Construct the relationship between output y(t) and control w(t), namely: Where q(t)=-Δb1y(t)-Δb2y(t)+Δb2w(t)+Δb2q2(t)+q1(t) is the equivalent perturbation; 4) Design a three-degree-of-freedom controller, including a feedforward controller, a PID feedback controller, and an equivalent disturbance compensation controller; The control input w(t) of the three-degree-of-freedom controller is as follows: w(t)=w FF (t)+w FB (t)+w DC (t) (5) Where w FF (t) is the feedforward controller input, w FB (t) is the input of PID feedback controller, w DC (t) is the input of the equivalent disturbance compensation controller; The piezoelectric driven cutting state control unit with force / torque feedback establishes a mapping relationship between cutting force and cutting parameters, and constructs a master-slave contact force feedback control strategy. After the mapping relationship between cutting force and cutting parameters is established, a force / torque feedback method is used to perform constant cutting force or constant cutting depth feeding, thereby realizing the master-slave contact force feedback control strategy; The vibration feedback piezoelectric driven cutting state control unit stores a vibration prediction model and a chatter prediction analysis model; The vibration prediction model and the chatter prediction analysis model are used to predict the vibration and chatter of the intelligent tool (1) respectively; The vibration feedback piezoelectric driven cutting state control unit actively controls the relative vibration between the intelligent tool and the workpiece according to vibration and chatter, and compensates for the dynamic stiffness under different cutting conditions.

2. The sensing, computing and control integrated intelligent tool system according to claim 1, characterized in that: The intelligent tool database module includes a cutting parameter and workpiece material query unit and an intelligent tool full working condition data storage and comparison unit; The cutting parameter and workpiece material query unit is used to query the cutting parameters corresponding to the workpiece material.

3. The sensing, computing and control integrated intelligent tool system according to claim 2, characterized in that: When the workpiece material being queried has been used, if the intelligent tool (1) has a response time value for an abnormal working condition of chipping or breaking of the blade, the alarm light of the intelligent tool (1) is permanently on, and after resetting, the workpiece material is re-entered or selected and the corresponding cutting parameters are given; If there is no response time value for the abnormal working condition of the intelligent tool (1) chipping or breaking, the operating condition data is read; the operating condition data includes cutting force and acceleration nonlinear errors, frequency response and cross interference errors, intelligent tool (1) wear online measurement error, and intelligent tool wear compensation positioning accuracy value; When the queried workpiece material is not used, the query operation stops.

4. The sensing, computing and control integrated intelligent tool system according to claim 2, characterized in that: The intelligent tool full working condition data storage and comparison unit is used to store cutting force and vibration data of the intelligent tool (1) during the machining process; When the tool wear high-precision timely in-situ compensation and state control module is completed, the intelligent tool full-working condition data storage and comparison unit also reads the cutting force and acceleration nonlinear errors, frequency response and cross-interference errors, intelligent tool wear online measurement errors and intelligent tool wear compensation positioning accuracy values, and compares them with historical records, retaining smaller error values ​​and smaller intelligent tool wear compensation positioning accuracy values.

5. The sensing, computing and control integrated intelligent tool system according to claim 1, characterized in that: The cutting force and vibration data perception module includes a cutting process force and vibration signal analysis unit and a cutting force and acceleration sensor state monitoring unit; The cutting process force and vibration signal analysis unit is used to select the arrangement position of the force sensor and the acceleration sensor; The force sensor and acceleration sensor are both arranged at the positions of the tool rod and the tool handle of the intelligent tool (1), provided that the force and acceleration signals meet the requirements that the nonlinear error is less than or equal to 0.3% FS and the frequency response is greater than or equal to 30 kHz; FS represents full scale; The cutting force and acceleration sensor condition monitoring unit stores a multi-dimensional force and acceleration cross-coupling analytical model; The cutting force and acceleration sensor state monitoring unit inputs the data monitored by the force sensor and the acceleration sensor into the multi-dimensional force and acceleration cross-coupling analytical model, suppresses the cross interference of the sensors, and processes the data to obtain the cutting force and vibration data; The steps to establish the multi-dimensional force and acceleration cross-coupling analytical model include: 1) The design coordinates of the acceleration sensor coincide with the setting coordinates of the intelligent tool; 2) The design coordinates of the force sensor coincide with the design coordinates of the acceleration sensor; 3) With the help of acceleration information, the coupling of inertial force to the force sensor is separated, and a multi-dimensional force and acceleration cross-coupling analytical model is established through iterative decoupling.

6. The sensing, computing and control integrated intelligent tool system according to claim 1, characterized in that: The high-precision in-situ real-time monitoring theory and calculation module of the tool status includes a tool dynamic evolution multi-source sensor information mapping unit, an abnormal feature recognition tool damage monitoring unit and a transfer learning multi-condition tool wear prediction unit; The tool dynamic evolution multi-source sensor information mapping unit is used to monitor the tool multi-dimensional force and acceleration data characteristic data; The tool breakage monitoring unit for abnormal feature recognition builds a multi-blade tool state change cloud map based on multi-dimensional force and acceleration data, and identifies the breakage and chipping of each tooth of the tool. If the response time of the intelligent tool chipping and tool breakage abnormal working condition is less than or equal to t, it enters the multi-condition tool wear prediction unit of transfer learning. Otherwise, it continues to build the multi-blade tool state change cloud map; The multi-working condition tool wear prediction unit of the transfer learning stores a transfer learning model for tool wear; The transfer learning model for tool wear is used to predict intelligent tool wear values.

7. The sensing, computing and control integrated intelligent tool system according to claim 6, characterized in that: The transfer learning model for tool wear is trained by a wear value smoothing method based on wear curve fitting; The extracted feature point x on the wear curve i obeys the normal distribution with a mean of 0 and a variance of The probability density function is Probability density function As shown below: Where X is a random vector in N-dimensional space, f(h) is the principal curve of X, ||Xf(h)|| is the Euclidean distance between X and the mapping point on the principal curve, l i is the curve length, is the variance of X and the principal curve; δ i is the standard deviation of X from the principal curve.

8. The sensing, computing and control integrated intelligent tool system according to claim 1, characterized in that: Also included is a charge amplifier (2); The charge amplifier (2) amplifies the vibration and cutting force data of the intelligent tool (1) and transmits the data to the data acquisition card (4).

9. The sensing, computing and control integrated intelligent tool system according to claim 1, characterized in that: Also included is a signal line (3); The intelligent tool (1) and the charge amplifier (2), the data acquisition card (4) and the numerical control system (5), and the charge amplifier (2) and the data acquisition card (4) perform data transmission via a signal line (3).

Images