Adjustable machining grinding equipment

Abstract

The invention discloses adjustable machining grinding equipment, and relates to the technical field of machining. According to the equipment, a multi-dimensional adjusting mechanism is innovatively integrated, a sliding block is driven through a rotatable double-thread lead screw, stepless adjustment of the grinding radius is achieved, and the equipment is suitable for machining of the inner diameters of workpieces of different sizes; and in combination with a dynamic balance weight system, vibration in high-speed grinding is effectively inhibited, and the stability is improved. Meanwhile, a pneumatic-servo cooperative control system is adopted, stepless switching of the grinding pressure in the four stages of light pressure, medium pressure, heavy pressure and fine grinding is supported, the revolution-rotation speed ratio of the grinding disc to the workpiece is synchronously regulated and controlled, and the uniformity of the grinding track is optimized. In addition, the equipment is provided with a PLC program control module and a touch screen interface, 30 sets of technological parameters can be preset and stored, one-key calling is achieved, and the machining efficiency and consistency of complex parts are remarkably improved. According to the design, the adaptability of the equipment to sapphire substrates, gears and hard alloy special-shaped high-hardness workpieces is greatly expanded.

Description

Technical Field

[0001] This invention relates to the field of machining technology, and in particular to an adjustable machining and grinding device. Background Technology

[0002] In the field of machining, grinding equipment is the core equipment for achieving high-precision surface treatment, especially for difficult-to-machine materials such as cemented carbide, ceramics, and precision bearings. Grinding accuracy directly determines the service life and performance stability of the workpiece. Current mainstream equipment mainly adjusts the grinding gap or pressure through mechanical structures, such as using lead screw mechanisms to adjust the grinding disc spacing and variable frequency motors to control the rotation speed. However, when dealing with irregularly shaped workpieces or high-hardness materials, problems such as insufficient adjustment flexibility, poor dynamic stability, and difficulty in maintaining accuracy are still prevalent. Existing technologies largely rely on operator experience to manually adjust parameters, making it difficult to achieve coordinated optimization of grinding pressure and motion trajectory, resulting in low processing consistency. Especially under high-speed conditions, vibration can easily cause surface scratches or dimensional deviations.

[0003] The typical scheme of the "Adjustable Polishing Device" disclosed in Chinese Utility Model Patent CN2848470Y consists of a support body and a polishing machine installed on it. The support body is a horizontal support (5) mounted on two parallel vertical supports (1). Both the horizontal support (5) and the vertical supports (1) are equipped with a horizontal lead screw (4) and a vertical lead screw (2). The vertical supports (1) are equipped with a vertical adjustment handle (6) for adjusting the movement of the horizontal support (5). Similarly, the horizontal support (5) is equipped with a horizontal adjustment handle (7) for adjusting the movement of the polishing machine (3) on the horizontal support (5). This utility model solves the problem of not being able to polish and grind large areas, and it is very convenient to adjust the direction and position of the polishing, with ideal results.

[0004] Based on the existing technologies mentioned above, adjustable grinding equipment still faces three key problems: Insufficient dynamic accuracy and stability: Mechanical adjustment mechanisms (such as lead screws and gears) are prone to thermal deformation and wear under varying working conditions, and have weak vibration suppression capabilities when rotating at high speeds, resulting in grinding pressure fluctuations (typical value > ±8%), which affect surface uniformity; Low level of intelligence: Parameter adjustment relies on manual preset, and it is impossible to dynamically optimize the grinding trajectory and pressure distribution according to the real-time shape of the workpiece (such as edge collapse and thickness deviation), resulting in a scrap rate of up to 18% for irregularly shaped parts. High maintenance costs: Traditional dynamic balancing uses mechanical counterweights, which take more than 3 seconds to adjust each time. Furthermore, material intrusion into the drive mechanism after the sealing structure fails will accelerate bearing wear, resulting in a mean time between failures (MTBF) of less than 500 hours.

[0005] Therefore, there is an urgent need for a new type of grinding equipment that integrates a high-response adjustment mechanism, online dynamic balance compensation, and intelligent parameter closed-loop control to meet the needs of precision manufacturing for micron-level consistent processing. Summary of the Invention

[0006] In view of the aforementioned existing problems, the present invention is proposed.

[0007] Therefore, the present invention provides an adjustable machining and grinding device to solve the problem of adjusting existing grinding devices. It suffers from poor flexibility, narrow applicability, and insufficient precision control.

[0008] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides an adjustable machining and grinding apparatus, comprising, The multi-degree-of-freedom adjustment mechanism consists of a planetary roller screw with a lead of 4mm and a repeatability of ±0.003mm and a cross roller guide, which drives the grinding disc to move along the X / Y / Z axes. The Z axis integrates a piezoelectric ceramic fine-tuning module with a resolution of 0.1μm. The dynamic balancing system consists of 32 sets of electromagnetic counterweight units distributed in a ring on the back side of the grinding disc. The centrifugal force deviation is detected in real time by Hall sensors, and the counterweight current is dynamically adjusted by a PID controller with a response time of ≤10ms. The composite drive module adopts a dual-rotor permanent magnet synchronous motor. The inner rotor is directly connected to the grinding disc to achieve stepless speed regulation from 0 to 10,000 r / min, while the outer rotor drives the workpiece tray to perform differential motion from 1 to 100 r / min through a harmonic reducer. The intelligent control system uses an embedded industrial PC to generate pressure-speed mapping curves based on a process database, and synchronously controls pneumatic servo valves and motor vector drives with a flow accuracy of ±0.05L / min via an EtherCAT bus.

[0009] As a preferred embodiment of the adjustable machining and grinding equipment described in this invention, the specific structural optimization of the multi-degree-of-freedom adjustment mechanism is as follows: The nut of the planetary roller screw adopts a double V-shaped raceway design, in which the main raceway bears the axial load and the secondary raceway bears the radial off-center load. The rollers are arranged in the double raceway at a 45° staggered angle. The roller assembly process adopts liquid nitrogen cold assembly process. First, the nut body is cooled to -196℃, and then the pre-ground GCr15 bearing steel rollers are embedded into the raceway at an alternating angle. After the temperature rises back to room temperature, an interference fit is formed, and the fit tolerance is controlled at H6 / p5 level. The preload maintenance mechanism consists of 8 sets of butterfly springs, each with a thickness of 1.5mm, stacked in a total of 6 pieces, forming a ring array. After the initial preload is applied to 1200N by a hydraulic servo cylinder, the spring group dynamically compensates for thermal deformation during operation, with the preload fluctuation range ≤±50N. The slider substrate of the cross roller guide is made of 38CrMoAl nitrided steel. The surface is coated with a 0.3mm thick WC-10Co hard alloy coating by laser cladding technology. The cladding power is 3.5kW, the scanning speed is 8mm / s, the coating microhardness is ≥1200HV, the friction coefficient is stable in the range of 0.08~0.10, and the straightness of the guide rail reaches 0.002mm / m after calibration by a laser interferometer.

[0010] As a preferred embodiment of the adjustable machining and grinding equipment described in this invention, the control process of the piezoelectric ceramic fine-tuning module is further refined: The stacked piezoelectric actuator consists of 120 PZT-8 piezoelectric ceramic sheets, each 0.1 mm thick, connected in series. It generates a linear displacement output of 0 to 80 μm within a driving voltage range of 0 to 150 V. The displacement resolution is controlled at 0.0012 μm by a 16-bit DAC module. The capacitive displacement sensor has its detection plate rigidly connected to the grinding disc. It adopts a differential measurement structure, with an excitation frequency of 1 MHz and a sensitivity of 0.2 V / μm. It acquires the Z-axis displacement signal of the grinding disc in real time and converts it into a digital quantity. The control process begins with an embedded industrial PC generating a feedforward voltage command based on a preset grinding pressure curve. This command is dynamically calculated based on the displacement-voltage transfer function of the piezoelectric actuator. Subsequently, a capacitive displacement sensor feeds back the actual displacement value at a sampling frequency of 5kHz. The feedforward voltage is then corrected by a digital PID controller with a proportional coefficient Kp=0.8, an integral time Ti=0.05s, and a derivative time Td=0.005s. Finally, a composite control signal is output to the high-voltage amplifier. When switching from coarse grinding to fine grinding, if the system detects that the pressure setpoint has dropped from 80 N / cm² to 30 N / cm², the feedforward module outputs a -42V step voltage to compensate for overshoot. At the same time, the PID controller resets the integral term to avoid integral saturation, so as to achieve a smooth pressure transition with a transition time of ≤20ms and an overshoot of <5%.

[0011] As a preferred embodiment of the adjustable machining and grinding equipment described in this invention, the execution process of the electromagnetic counterweight unit of the dynamic balancing system is specified as follows: The neodymium iron boron permanent magnets of the electromagnetic counterweight unit are fixed in the annular groove of the titanium alloy base on the back side of the grinding disc by high temperature epoxy resin. The 32 sets of electromagnetic counterweight units are distributed at equal angles. The excitation coil uses high-temperature resistant polyimide enameled wire with a diameter of 0.5mm and 120 turns, wound into a hollow cylindrical structure. The coil frame is a zirconia ceramic tube with a wall thickness of 0.8mm. When the counterweight adjustment is started, the Hall sensor embedded in the center of the coil, with a sensitivity of 1.2mV / G, collects the magnetic field vector generated by the centrifugal force in real time, and converts it into a digital signal by the 24-bit ADC module at a sampling rate of 10kHz. The signal processing unit performs a 1024-point FFT analysis on the raw data, extracting the vibration amplitude and phase angle of 1 to 3 times the fundamental frequency. When the fundamental frequency amplitude is detected to exceed 5 μm, the PID controller Kp=1.5, Ki=0.2, Kd=0.02, and calculates the compensation current value based on the phase angle, ranging from 0 to 5A. This compensation current is applied to the excitation coil via the PWM drive module. The reverse electromagnetic field generated by the current couples with the permanent magnet to form a maximum anti-vibration torque of 15 N·m, which cancels out the imbalance. The system has a pre-stored static imbalance compensation curve: below 3000 r / min, a linear compensation mode is used with a current gain of 0.8 A / g·mm; above 3000 r / min, it switches to quadratic function compensation with a current gain of 1.2 A / g·mm - 0.0003 A / rpm². At the same time, the coil temperature is monitored by thermocouples and the current coefficient is dynamically corrected, with a temperature compensation rate of -0.3% / ℃.

[0012] As a preferred embodiment of the adjustable machining and grinding equipment described in this invention, the dual-rotor permanent magnet synchronous motor of the composite drive module undergoes structural refinement and coordinated control. The inner rotor adopts a Halbach array magnet arrangement, consisting of 16 N50SH neodymium iron boron magnets bonded sequentially to the SMC soft magnetic composite matrix with a deflection angle of 22.5° per pole, forming a unilateral enhanced magnetic field, so that the air gap magnetic flux density reaches 1.35T. The outer rotor is made of 0.1mm thick Fe-Si-Cr soft magnetic alloy sheets stacked together. 36 wedge-shaped grooves are opened on the surface of the pole shoes to suppress eddy current losses. Distributed temperature sensors are embedded in the rotor yoke. The stator winding adopts a segmented structure, with each phase consisting of 6 independent sector coils with a wire diameter of 1.2mm and 12 turns per coil. A 0.8mm cooling channel is reserved between the coils, and copper heat-conducting fins with a thickness of 0.3mm are welded to the ends of each coil. The cooling system consists of a two-stage constant-temperature circulation device. The first-stage circulation cools deionized water to 25±0.5℃ via a semiconductor refrigeration module. The second-stage circulation injects coolant into the annular main pipe on the back of the stator through a micro turbine pump at a flow rate of 12L / min, and then distributes it to each coil flow channel through 12 stainless steel capillary tubes. The outlet water temperature monitoring accuracy is ±0.3℃. The motor control adopts a dual closed-loop strategy: the inner loop uses sensorless position detection based on high-frequency signal injection method with an angular resolution of 0.01°. The outer loop achieves stepless speed regulation of the inner rotor from 0 to 10,000 r / min through magnetic field orientation control (FOC). At the same time, the outer rotor drives the workpiece tray to move at a differential speed by a harmonic reducer. The torque coupling between the inner and outer rotors is dynamically compensated by a load observer to ensure that the speed fluctuation is ≤±0.05%.

[0013] As a preferred embodiment of the adjustable machining and grinding equipment described in this invention, the differential motion control process of the composite drive module is refined as follows: The coordinated motion of the inner and outer rotors of the dual-rotor permanent magnet synchronous motor is achieved through an electronic gearbox. The specific execution process is as follows: the motion controller receives the speed ratio command set by the process parameters, generates the target speed curve for the inner rotor based on an S-curve acceleration / deceleration algorithm, limiting the acceleration to within 500 rpm / s², while the target speed of the outer rotor is determined according to the formula ω. out =k·ω in +b is calculated dynamically, where k is the speed ratio coefficient, and ω in ω represents the actual rotational speed of the inner rotor, where b is the base rotational speed (1~10 r / min); ω represents the actual rotational speed of the inner rotor. in The rotational speed is detected in real time by a rotary transformer. The detection signal is converted into a digital quantity by an RDC decoding chip. The rotational speed of the outer rotor is collected by a circular grating encoder on the output shaft of the harmonic reducer, with 50,000 pulses per revolution. Speed ​​error compensation employs a feedforward-feedback composite control: the feedforward channel is based on the disturbance torque predicted by the load observer, with a sampling period of 100μs, and suppresses speed fluctuations through pre-compensation current; the feedback channel uses a dual closed-loop PID, with speed loop Kp=0.6 and integral time of 0.1s, and position loop Kp=12 and integral time of 0.05s; when the speed ratio switches from coarse grinding mode k=0.4 to fine grinding mode k=0.9, the system first smoothly transitions the k value with a time constant of 20ms, while dynamically adjusting the b value to compensate for inertial delay, and the outer rotor acceleration is synchronously limited to 30rpm / s²; the control signal is transmitted to the dual-channel vector driver via the EtherCAT bus, the inner rotor driver uses space vector PWM with a switching frequency of 8kHz, and the outer rotor driver injects a third harmonic to enhance low-speed torque with a carrier frequency of 4kHz.

[0014] In a preferred embodiment of the adjustable machining and grinding equipment described in this invention, the parameter calling and switching control of the process database is specifically implemented as follows: The process database stores 30 sets of standard grinding formulas. Each set of formulas includes a pressure gradient curve, four stages of coarse grinding / semi-fine grinding / fine grinding / ultra-fine grinding, a speed ratio dynamic function, abrasive formula code, and processing time threshold. When parameters are called, the embedded industrial PC automatically matches the basic parameters based on the workpiece material code, and then adds the correction coefficient according to the workpiece morphology scanning data; the abrasive supply system is equipped with a four-channel precision metering pump, which accurately mixes diamond micro powder with a concentration of 5~30%, cubic boron nitride suspension with a concentration of 3~15%, cerium oxide polishing liquid and pH adjuster according to the formula code, with a mixing accuracy of ±0.5ml / min. The parameter switching control adopts a three-level fuzzy PID strategy: In the first level, during the transition from coarse grinding to semi-fine grinding, the pressure changes from 80 N / cm² to 50 N / cm², with a pressure change rate of 2.5 N / cm²·s as the constraint. The proportional gain is adaptively adjusted, and Kp linearly decreases from 0.8 to 0.3. In the second level, during the transition from semi-fine grinding to fine grinding, the speed ratio k changes from 0.5 to 0.9. A feedforward compensation algorithm is used, based on the speed difference, to preload compensation torque, ΔT=0.05·J·dω / dt, where J is the identified value of rotational inertia. In the third level, when the ultra-fine grinding stage starts, the in-situ measurement system is synchronously triggered to evaluate the surface quality. If the roughness Sa>0.02μm, the fine grinding time is extended by 15% and the speed ratio is increased to 0.95. If Sa≤0.02μm, the original parameters are maintained. All parameter switching processes are displayed in real time on the HMI interface, and the actual execution curve and theoretical deviation value are recorded. The pressure fluctuation is <±1.5%, and the speed ratio error is <±0.01.

[0015] As a preferred embodiment of the adjustable machining and grinding equipment described in this invention, the execution flow of the in-situ measurement system is specifically implemented as follows: When the grinding stage switching signal is triggered, the pneumatic lifting device, with a cylinder diameter of Φ40mm and a stroke of 25mm, pushes the ceramic push rod with a pressure of 0.5MPa and an elastic modulus of 300GPa, and lifts the grinding disc by 1.5±0.02mm within 120ms. During the separation process, the piezoelectric ceramic fine-tuning module synchronously outputs reverse displacement to compensate for mechanical vibration. A white light interferometer, with a wavelength of 460 nm and a vertical scanning range of 500 μm, is moved to the center of the workpiece via a three-axis precision displacement stage and scans along a spiral path with a radius from 0 to R. max R max At the maximum radius, with an angular velocity of 10 rad / s, surface topography data was collected with a scan point spacing of 2.5 μm and a collection time of ≤0.8 s per frame. The topography processing unit reconstructs the three-dimensional surface based on the phase-shifting interferometry algorithm and extracts key parameters: the Sa value of the central region with an area diameter of 2 mm, the edge collapse amount of the height difference within the outermost ring width of 0.3 mm, and the total thickness change TTV; if Sa > 0.05 μm, the system automatically extends the fine grinding time by 20% and increases the rotation speed ratio by 0.05; if Sa ≤ 0.05 μm, the acoustic emission monitoring process is activated. The acoustic emission sensor, with a resonant frequency of 150kHz and a sensitivity of 85dB, is attached to the back of the workpiece tray via a magnetic base. After the signal is filtered by a 100~500kHz bandpass filter and pre-amplified by 40dB, the characteristic parameters are analyzed in real time by the FPGA module: when a burst wave with an amplitude > 85dB and a duration > 50μs is detected, it is determined to be a crack propagation signal, and grinding is immediately stopped and an alarm is triggered. Measurement data and processing results are transmitted to the database via industrial Ethernet to update process parameters. At the same time, the HMI interface displays a three-dimensional topographic cloud map and defect location coordinates. The entire in-situ measurement cycle is controlled within 15 seconds.

[0016] As a preferred embodiment of the adjustable machining and grinding equipment described in this invention, the specific execution flow of the fault diagnosis module is refined as follows: When the equipment is started, a high-precision current sensor with a bandwidth of 0~10kHz collects the current signal of the three-phase winding of the dual-rotor motor in real time. Simultaneously, a MEMS accelerometer with a range of ±50g is installed on the radial and axial measuring points of the motor housing through a magnetic base. The signal processing unit acquires raw data at a sampling rate of 10kHz. After the current signal is filtered by an FIR low-pass filter with a cutoff frequency of 1.5kHz to eliminate switching noise, it is analyzed by FFT with a 4096-point Hanning window to extract the fundamental to 15th harmonic components. The vibration signal is decomposed by wavelet packet at 6 levels to extract the energy values ​​of eight sub-bands from 50 to 8000Hz. The diagnostic logic executes a three-step judgment: Step 1: Detect the third harmonic content of the current. If the increase is greater than 15% for five consecutive sampling cycles and accompanied by a 20% increase in vibration energy in the 100-200Hz range, it is judged as bearing wear, code F01, triggering a yellow warning and recommending bearing replacement within 72 hours; Step 2: Monitor the vibration spectrum. When the amplitude at the 800Hz frequency point exceeds three times the baseline value for 10 seconds, and the fifth harmonic fluctuation rate of the current is greater than 8%, it is judged as abrasive blockage, code F02, initiating the automatic cleaning program of the grinding disc, with a reverse flushing pressure of 0.8MPa for 15 seconds; Step 3: Integrate temperature data. If the stator winding temperature rise rate is greater than 1.5℃ / s and the vibration energy in the 200-400Hz range suddenly increases by 30%, it is judged as cooling failure, code F03, immediately reducing the speed to a safe value and closing the coolant valve; All diagnostic results are encapsulated into JSON format messages via the OPC UA protocol and pushed to the MES system production dashboard in real time. At the same time, the local HMI interface displays a fault location map and handling guide. The diagnostic response delay is ≤50ms, and the false alarm rate has been verified to be <0.8% after thousands of hours of testing.

[0017] In a preferred embodiment of the adjustable machining and grinding equipment described in this invention, the adaptive optimization module of the intelligent control system is specifically implemented as follows: During equipment operation, the embedded industrial PC collects a multi-dimensional status dataset every 5 minutes, including the displacement fluctuation σ of the piezoelectric fine-tuning module. d The electromagnetic counterweight unit current change rate di / dt, the dual rotor motor winding temperature gradient ΔT / Δt, and the in-situ measured Sa value sequence; Data preprocessing employed a sliding window filter with a window width of 10 periods and a cutoff frequency of 0.1 Hz. After noise removal, seven key features were extracted: displacement fluctuation σ. d >0.15μm is marked as characteristic F1, di / dt>2A / s is marked as F2, winding temperature rise rate>1.2℃ / s is marked as F3, Sa value increases for 3 consecutive times is marked as F4, vibration 800Hz frequency band energy increase>25% is marked as F5, current third harmonic continuously exceeds the standard is marked as F6, abrasive flow rate fluctuation>8% is marked as F7; A decision model based on the random forest algorithm was pre-trained with 50,000 samples. Real-time analysis of feature combinations was performed: when the combination [F1+F4+F6] was detected, the flatness of the grinding disc was determined to be deteriorated, and the disc dressing program was automatically activated, with a diamond dressing wheel feed of 0.005 mm / cycle and a dressing time of 120 s; when the combination [F2+F5+F7] appeared, the abrasive pipeline was determined to be blocked, and a three-stage pulse backflushing was executed, with pressure increasing in stages of 0.6 / 1.0 / 1.4 MPa, 5 pulses per stage; when the combination [F3+F6] lasted for 2 cycles, the bearing lubrication was determined to be ineffective, and a special grease was injected, with a dosage of 0.2 ml and a viscosity of ISO VG 32. All optimization operations are recorded in a blockchain-based process log, encrypted with SHA-256. The Sa value is re-evaluated after each optimization: if the Sa decrease is less than 10%, the expert system is activated to intervene and call the optimization parameters of similar working conditions in the case library to refine the grinding pressure or extend the time. After processing 20 workpieces, the offline calibration module automatically compares the key dimensions, flatness / cylindricity / roughness of the first and last workpieces, generates an accuracy decay curve, and reverses the weight coefficients of the decision model. The correction amount ΔW = 0.05 × |Δaccuracy|, realizing closed-loop self-evolutionary control.

[0018] In a second aspect, the present invention provides a computer device including a memory and a processor, wherein the memory stores a computer program, and the computer program, when executed by the processor, implements an adjustable machining and grinding device as described in the first aspect of the present invention.

[0019] Thirdly, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements an adjustable machining and grinding apparatus as described in the first aspect of the present invention.

[0020] The beneficial effects of this invention are: This invention achieves nanoscale grinding pressure control (accuracy ±0.5N) through the synergistic innovation of a multi-degree-of-freedom adjustment mechanism and a piezoelectric ceramic fine-tuning module. Combined with the planetary roller screw double V-shaped raceway design and laser-clad WC-10Co guide rail coating (friction coefficient 0.08), it effectively eliminates the accuracy drift caused by backlash and thermal deformation in traditional screw drives, ensuring that the flatness of the carbide cutting edge is stably controlled within ±0.0005mm, improving accuracy by 4 times compared to existing equipment. At the same time, the electromagnetic dynamic balancing system (response time ≤10ms) replaces the mechanical counterweight with an active electromagnetic counterweight. By analyzing the vibration spectrum in real time using Fourier transform and outputting a reverse compensation current, it completely solves the chatter problem caused by centrifugal force deviation under high-speed conditions (10,000r / min). The measured surface roughness fluctuation of aerospace titanium alloy blades has been reduced from 35% to 2%, significantly improving the pass rate of thin-walled parts processing.

[0021] Furthermore, the deep integration of the dual-rotor permanent magnet synchronous motor and the intelligent control system breaks through the bottleneck of traditional differential motion. The inner rotor Halbach magnet array (air gap magnetic flux density 1.35T) combined with the outer rotor harmonic reduction drive achieves stepless adjustment of the speed ratio from 0.1 to 1.0 and a synchronization accuracy of ±0.03%. Combined with the S-curve acceleration and deceleration algorithm to optimize the uniformity of the grinding trajectory, the surface roughness distribution of the automotive turbocharger impeller flow channel achieves a uniformity of 94.7%, and the processing cycle is shortened by 25%. More importantly, the process database and the in-situ measurement system form a closed-loop feedback. The white light interferometer (vertical resolution 0.1nm) completes the three-dimensional morphology scan of the workpiece within 15 seconds. Based on the real-time Sa value, the fine grinding time and speed ratio are dynamically adjusted (e.g., the time is extended by 20% when Sa>0.05μm). The acoustic emission module (capture sensitivity 85dB) accurately identifies micro-crack signals, reducing the scrap rate of batch processing of irregularly shaped workpieces from 18% to 0.3%.

[0022] Ultimately, the fault diagnosis and adaptive optimization module constructs a predictive maintenance system. The random forest algorithm (50 decision trees) analyzes 7 features (such as displacement fluctuation σ) every 5 minutes. d(Exceeding 0.15μm and 3rd harmonic current limits), the system provides a 72-hour advance warning of bearing wear and automatically triggers a repair program. Combined with blockchain process logs to track the precision decay curve, the system ensures that after 1000 hours of continuous operation, the critical dimension drift is ≤±0.0005mm, the mean time between failures (MTBF) is extended from 500 hours to 3500 hours, and maintenance costs are reduced by 70%. In the manufacturing of high-value-added parts such as aerospace gears and artificial joints, the system achieves an annual comprehensive efficiency (OEE) of 89.5% for a single unit, providing a breakthrough solution for domestically produced high-end equipment in the precision manufacturing field. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. 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.

[0024] Figure 1 This is a block diagram of an adjustable machining and grinding device in Example 1. Detailed Implementation

[0025] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0026] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0027] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0028] Example 1, referring to Figure 1 This is the first embodiment of the present invention, which provides an adjustable machining and grinding device, comprising: The multi-degree-of-freedom adjustment mechanism consists of a planetary roller screw with a lead of 4mm and a repeatability of ±0.003mm and a cross roller guide, which drives the grinding disc to move along the X / Y / Z axes. The Z axis integrates a piezoelectric ceramic fine-tuning module with a resolution of 0.1μm. The dynamic balancing system consists of 32 sets of electromagnetic counterweight units distributed in a ring on the back side of the grinding disc. The centrifugal force deviation is detected in real time by Hall sensors, and the counterweight current is dynamically adjusted by a PID controller with a response time of ≤10ms. The composite drive module adopts a dual-rotor permanent magnet synchronous motor. The inner rotor is directly connected to the grinding disc to achieve stepless speed regulation from 0 to 10,000 r / min, while the outer rotor drives the workpiece tray to perform differential motion from 1 to 100 r / min through a harmonic reducer. The intelligent control system uses an embedded industrial PC to generate pressure-speed mapping curves based on a process database, and synchronously controls pneumatic servo valves and motor vector drives with a flow accuracy of ±0.05L / min via an EtherCAT bus.

[0029] Specific structural optimization of the multi-degree-of-freedom adjustment mechanism: The nut of the planetary roller screw adopts a double V-shaped raceway design, in which the main raceway bears the axial load and the secondary raceway bears the radial off-center load. The rollers are arranged in the double raceway at a 45° staggered angle. The roller assembly process adopts liquid nitrogen cold assembly process. First, the nut body is cooled to -196℃, and then the pre-ground GCr15 bearing steel rollers are embedded into the raceway at an alternating angle. After the temperature rises back to room temperature, an interference fit is formed, and the fit tolerance is controlled at H6 / p5 level. The preload maintenance mechanism consists of 8 sets of butterfly springs, each with a thickness of 1.5mm, stacked in a total of 6 pieces, forming a ring array. After the initial preload is applied to 1200N by a hydraulic servo cylinder, the spring group dynamically compensates for thermal deformation during operation, with the preload fluctuation range ≤±50N. The slider substrate of the cross roller guide is made of 38CrMoAl nitrided steel. The surface is coated with a 0.3mm thick WC-10Co hard alloy coating by laser cladding technology. The cladding power is 3.5kW, the scanning speed is 8mm / s, the coating microhardness is ≥1200HV, the friction coefficient is stable in the range of 0.08~0.10, and the straightness of the guide rail reaches 0.002mm / m after calibration by a laser interferometer.

[0030] Further refine the control process of the piezoelectric ceramic fine-tuning module: The stacked piezoelectric actuator consists of 120 PZT-8 piezoelectric ceramic sheets, each 0.1 mm thick, connected in series. It generates a linear displacement output of 0 to 80 μm within a driving voltage range of 0 to 150 V. The displacement resolution is controlled at 0.0012 μm by a 16-bit DAC module. The capacitive displacement sensor has its detection plate rigidly connected to the grinding disc. It adopts a differential measurement structure, with an excitation frequency of 1 MHz and a sensitivity of 0.2 V / μm. It acquires the Z-axis displacement signal of the grinding disc in real time and converts it into a digital quantity. The control process begins with an embedded industrial PC generating a feedforward voltage command based on a preset grinding pressure curve. This command is dynamically calculated based on the displacement-voltage transfer function of the piezoelectric actuator. Subsequently, a capacitive displacement sensor feeds back the actual displacement value at a sampling frequency of 5kHz. The feedforward voltage is then corrected by a digital PID controller with a proportional coefficient Kp=0.8, an integral time Ti=0.05s, and a derivative time Td=0.005s. Finally, a composite control signal is output to the high-voltage amplifier. When switching from coarse grinding to fine grinding, if the system detects that the pressure setpoint has dropped from 80 N / cm² to 30 N / cm², the feedforward module outputs a -42V step voltage to compensate for overshoot. At the same time, the PID controller resets the integral term to avoid integral saturation, so as to achieve a smooth pressure transition with a transition time of ≤20ms and an overshoot of <5%.

[0031] The execution process of the electromagnetic counterweight unit in the dynamic balancing system is specified as follows: The neodymium iron boron permanent magnets of the electromagnetic counterweight unit are fixed in the annular groove of the titanium alloy base on the back side of the grinding disc by high temperature epoxy resin. The 32 sets of electromagnetic counterweight units are distributed at equal angles. The excitation coil uses high-temperature resistant polyimide enameled wire with a diameter of 0.5mm and 120 turns, wound into a hollow cylindrical structure. The coil frame is a zirconia ceramic tube with a wall thickness of 0.8mm. When the counterweight adjustment is started, a Hall sensor embedded in the center of the coil with a sensitivity of 1.2mV / G collects the magnetic field vector generated by the centrifugal force in real time, which is converted into a digital signal by a 24-bit ADC module at a sampling rate of 10kHz. The signal processing unit performs a 1024-point FFT analysis on the raw data, extracting the vibration amplitude and phase angle of 1 to 3 times the fundamental frequency. When the fundamental frequency amplitude is detected to exceed 5 μm, the PID controller Kp=1.5, Ki=0.2, Kd=0.02, and calculates the compensation current value based on the phase angle, ranging from 0 to 5A. This compensation current is applied to the excitation coil via the PWM drive module. The reverse electromagnetic field generated by the current couples with the permanent magnet to form a maximum anti-vibration torque of 15 N·m, which cancels out the imbalance. The system has a pre-stored static imbalance compensation curve: below 3000 r / min, a linear compensation mode is used with a current gain of 0.8 A / g·mm; above 3000 r / min, it switches to quadratic function compensation with a current gain of 1.2 A / g·mm - 0.0003 A / rpm². At the same time, the coil temperature is monitored by thermocouples and the current coefficient is dynamically corrected, with a temperature compensation rate of -0.3% / ℃.

[0032] Structural refinement and coordinated control of a dual-rotor permanent magnet synchronous motor with a composite drive module: The inner rotor adopts a Halbach array magnet arrangement, consisting of 16 N50SH neodymium iron boron magnets bonded sequentially to the SMC soft magnetic composite matrix with a deflection angle of 22.5° per pole, forming a unilateral enhanced magnetic field, so that the air gap magnetic flux density reaches 1.35T. The outer rotor is made of 0.1mm thick Fe-Si-Cr soft magnetic alloy sheets stacked together. 36 wedge-shaped grooves are opened on the surface of the pole shoes to suppress eddy current losses. Distributed temperature sensors are embedded in the rotor yoke. The stator winding adopts a segmented structure, with each phase consisting of 6 independent sector coils with a wire diameter of 1.2mm and 12 turns per coil. A 0.8mm cooling channel is reserved between the coils, and copper heat-conducting fins with a thickness of 0.3mm are welded to the ends of each coil. The cooling system consists of a two-stage constant temperature circulation device. The first-stage circulation cools the deionized water to 25±0.5℃ through the semiconductor refrigeration module. The second-stage circulation injects the coolant into the annular main pipe on the back of the stator through a micro turbine pump at a flow rate of 12L / min. The coolant is then distributed to each coil flow channel through 12 stainless steel capillary tubes. The outlet water temperature monitoring accuracy is ±0.3℃. The motor control adopts a dual closed-loop strategy: the inner loop is a sensorless position detection based on high-frequency signal injection with an angular resolution of 0.01°, and the outer loop achieves stepless speed regulation of the inner rotor from 0 to 10,000 r / min through field-oriented control (FOC). At the same time, the outer rotor is driven by a harmonic reducer to make differential motion of the workpiece tray. The torque coupling between the inner and outer rotors is dynamically compensated by the load observer to ensure that the speed fluctuation is ≤ ±0.05%.

[0033] The differential motion control process of the composite drive module is refined as follows: The coordinated motion of the inner and outer rotors of the dual-rotor permanent magnet synchronous motor is achieved through an electronic gearbox. The specific execution process is as follows: the motion controller receives the speed ratio command set by the process parameters, generates the target speed curve for the inner rotor based on an S-curve acceleration / deceleration algorithm, limiting the acceleration to within 500 rpm / s², while the target speed of the outer rotor is determined according to the formula ω. out =k·ω in +b is calculated dynamically, where k is the speed ratio coefficient, and ω in ω represents the actual rotational speed of the inner rotor, where b is the base rotational speed (1~10 r / min); ω represents the actual rotational speed of the inner rotor. in The rotational speed is detected in real time by a rotary transformer. The detection signal is converted into a digital quantity by an RDC decoding chip. The rotational speed of the outer rotor is collected by a circular grating encoder on the output shaft of the harmonic reducer, with 50,000 pulses per revolution. Speed ​​error compensation employs a feedforward-feedback composite control: the feedforward channel is based on the disturbance torque predicted by the load observer, with a sampling period of 100μs, and suppresses speed fluctuations through pre-compensation current; the feedback channel uses a dual closed-loop PID, with speed loop Kp=0.6 and integral time of 0.1s, and position loop Kp=12 and integral time of 0.05s; when the speed ratio switches from coarse grinding mode k=0.4 to fine grinding mode k=0.9, the system first smoothly transitions the k value with a time constant of 20ms, while dynamically adjusting the b value to compensate for inertial delay, and the outer rotor acceleration is synchronously limited to 30rpm / s²; the control signal is transmitted to the dual-channel vector driver via the EtherCAT bus, the inner rotor driver uses space vector PWM with a switching frequency of 8kHz, and the outer rotor driver injects a third harmonic to enhance low-speed torque with a carrier frequency of 4kHz.

[0034] The specific implementation of parameter calling and switching control in the process database is as follows: The process database stores 30 sets of standard grinding formulas. Each formula includes a pressure gradient curve, four stages (rough grinding, semi-fine grinding, fine grinding, and ultra-fine grinding), a dynamic function of rotational speed ratio, abrasive formula code, and processing time threshold. When parameters are called, the embedded industrial PC automatically matches the basic parameters based on the workpiece material code, and then adds correction coefficients based on the workpiece morphology scanning data. The abrasive supply system is equipped with a four-channel precision metering pump, which accurately mixes diamond micro powder (5-30%), cubic boron nitride suspension (3-15%), cerium oxide polishing slurry, and pH adjuster according to the formula code, with a mixing accuracy of ±0.5 ml / min. The parameter switching control adopts a three-level fuzzy PID strategy: In the first level, during the transition from coarse grinding to semi-fine grinding, the pressure changes from 80 N / cm² to 50 N / cm², with a pressure change rate of 2.5 N / cm²·s as the constraint. The proportional gain is adaptively adjusted, and Kp linearly decreases from 0.8 to 0.3. In the second level, during the transition from semi-fine grinding to fine grinding, the speed ratio k changes from 0.5 to 0.9. A feedforward compensation algorithm is used, based on the speed difference, to preload compensation torque, ΔT=0.05·J·dω / dt, where J is the identified value of rotational inertia. In the third level, when the ultra-fine grinding stage starts, the in-situ measurement system is synchronously triggered to evaluate the surface quality. If the roughness Sa>0.02μm, the fine grinding time is extended by 15% and the speed ratio is increased to 0.95. If Sa≤0.02μm, the original parameters are maintained. All parameter switching processes are displayed in real time on the HMI interface, and the actual execution curve and theoretical deviation value are recorded. The pressure fluctuation is <±1.5%, and the speed ratio error is <±0.01.

[0035] The execution process of the in-situ measurement system is implemented in detail as follows: When the grinding stage switching signal is triggered, the pneumatic lifting device, with a cylinder diameter of Φ40mm and a stroke of 25mm, pushes the ceramic push rod with a pressure of 0.5MPa and an elastic modulus of 300GPa, and lifts the grinding disc by 1.5±0.02mm within 120ms. During the separation process, the piezoelectric ceramic fine-tuning module synchronously outputs reverse displacement to compensate for mechanical vibration. A white light interferometer, with a wavelength of 460 nm and a vertical scanning range of 500 μm, is moved to the center of the workpiece via a three-axis precision displacement stage and scans along a spiral path with a radius from 0 to R. max R max At the maximum radius, with an angular velocity of 10 rad / s, surface topography data was collected with a scan point spacing of 2.5 μm and a collection time of ≤0.8 s per frame. The topography processing unit reconstructs the three-dimensional surface based on the phase-shifting interferometry algorithm and extracts key parameters: the Sa value of the central region with an area diameter of 2 mm, the edge collapse amount of the height difference within the outermost ring width of 0.3 mm, and the total thickness change TTV; if Sa > 0.05 μm, the system automatically extends the fine grinding time by 20% and increases the rotation speed ratio by 0.05; if Sa ≤ 0.05 μm, the acoustic emission monitoring process is activated. The acoustic emission sensor, with a resonant frequency of 150kHz and a sensitivity of 85dB, is attached to the back of the workpiece tray via a magnetic base. After the signal is filtered by a 100~500kHz bandpass filter and pre-amplified by 40dB, the characteristic parameters are analyzed in real time by the FPGA module: when a burst wave with an amplitude > 85dB and a duration > 50μs is detected, it is determined to be a crack propagation signal, and grinding is immediately stopped and an alarm is triggered. Measurement data and processing results are transmitted to the database via industrial Ethernet to update process parameters. At the same time, the HMI interface displays a three-dimensional topographic cloud map and defect location coordinates. The entire in-situ measurement cycle is controlled within 15 seconds.

[0036] The specific execution flow of the fault diagnosis module is detailed as follows: When the equipment is started, a high-precision current sensor with a bandwidth of 0~10kHz collects the current signal of the three-phase winding of the dual-rotor motor in real time. Simultaneously, a MEMS accelerometer with a range of ±50g is installed on the radial and axial measuring points of the motor housing through a magnetic base. The signal processing unit acquires raw data at a sampling rate of 10kHz. After the current signal is filtered by an FIR low-pass filter with a cutoff frequency of 1.5kHz to eliminate switching noise, it is analyzed by FFT with a 4096-point Hanning window to extract the fundamental to 15th harmonic components. The vibration signal is decomposed by wavelet packet at 6 levels to extract the energy values ​​of eight sub-bands from 50 to 8000Hz. The diagnostic logic executes a three-step judgment: Step 1: Detect the third harmonic content of the current. If the increase is greater than 15% for five consecutive sampling cycles and accompanied by a 20% increase in vibration energy in the 100-200Hz range, it is judged as bearing wear, code F01, triggering a yellow warning and recommending bearing replacement within 72 hours; Step 2: Monitor the vibration spectrum. When the amplitude at the 800Hz frequency point exceeds three times the baseline value for 10 seconds, and the fifth harmonic fluctuation rate of the current is greater than 8%, it is judged as abrasive blockage, code F02, initiating the automatic cleaning program of the grinding disc, with a reverse flushing pressure of 0.8MPa for 15 seconds; Step 3: Integrate temperature data. If the stator winding temperature rise rate is greater than 1.5℃ / s and the vibration energy in the 200-400Hz range suddenly increases by 30%, it is judged as cooling failure, code F03, immediately reducing the speed to a safe value and closing the coolant valve; All diagnostic results are encapsulated into JSON format messages via the OPC UA protocol and pushed to the MES system production dashboard in real time. At the same time, the local HMI interface displays a fault location map and handling guide. The diagnostic response delay is ≤50ms, and the false alarm rate has been verified to be <0.8% after thousands of hours of testing.

[0037] The adaptive optimization module of the intelligent control system is implemented in detail as follows: During equipment operation, the embedded industrial PC collects a multi-dimensional status dataset every 5 minutes, including the displacement fluctuation σ of the piezoelectric fine-tuning module. d The electromagnetic counterweight unit current change rate di / dt, the dual rotor motor winding temperature gradient ΔT / Δt, and the in-situ measured Sa value sequence; Data preprocessing employed a sliding window filter with a window width of 10 periods and a cutoff frequency of 0.1 Hz. After noise removal, seven key features were extracted: displacement fluctuation σ. d >0.15μm is marked as characteristic F1, di / dt>2A / s is marked as F2, winding temperature rise rate>1.2℃ / s is marked as F3, Sa value increases for 3 consecutive times is marked as F4, vibration 800Hz frequency band energy increase>25% is marked as F5, current third harmonic continuously exceeds the standard is marked as F6, abrasive flow rate fluctuation>8% is marked as F7; A decision model based on the random forest algorithm was pre-trained with 50,000 samples. Real-time analysis of feature combinations was performed: when the combination [F1+F4+F6] was detected, the flatness of the grinding disc was determined to be deteriorated, and the disc dressing program was automatically activated, with a diamond dressing wheel feed of 0.005 mm / cycle and a dressing time of 120 s; when the combination [F2+F5+F7] appeared, the abrasive pipeline was determined to be blocked, and a three-stage pulse backflushing was executed, with pressure increasing in stages of 0.6 / 1.0 / 1.4 MPa, 5 pulses per stage; when the combination [F3+F6] lasted for 2 cycles, the bearing lubrication was determined to be ineffective, and a special grease was injected, with a dosage of 0.2 ml and a viscosity of ISO VG 32. All optimization operations are recorded in a blockchain-based process log, encrypted with SHA-256. The Sa value is re-evaluated after each optimization: if the Sa decrease is less than 10%, the expert system is activated to intervene and call the optimization parameters of similar working conditions in the case library to refine the grinding pressure or extend the time. After processing 20 workpieces, the offline calibration module automatically compares the key dimensions, flatness / cylindricity / roughness of the first and last workpieces, generates an accuracy decay curve, and reverses the weight coefficients of the decision model. The correction amount ΔW = 0.05 × |Δaccuracy|, realizing closed-loop self-evolutionary control.

[0038] Example 2 is the second embodiment of the present invention. The working process of an adjustable machining and grinding device in this embodiment is as follows: After the equipment is started, the operator inputs the workpiece parameters (such as material GCr15, diameter Φ50±0.01mm) through the HMI interface of the intelligent control system. The embedded industrial PC automatically matches the grinding formula according to the process database (such as P07 group: rough grinding pressure 75N / cm², speed ratio k=0.35, fine grinding time 180s). The multi-degree-of-freedom adjustment mechanism acts first: the planetary roller screw drives the grinding disk to descend along the Z-axis to the initial gap position of 0.5mm. At this time, the piezoelectric ceramic fine adjustment module preloads 20N contact pressure; at the same time, the dynamic balance system starts self-check, the Hall sensor collects the static imbalance of the grinding disk (if the detected offset is >2g·mm), the PID controller outputs 3.2A current to the corresponding electromagnetic counterweight unit, and the initial balance compensation is completed within 8ms.

[0039] The composite drive module is then activated: the inner rotor of the dual-rotor permanent magnet synchronous motor accelerates to the roughing speed of 6000 r / min using an S-curve (acceleration limited to 500 rpm / s²), and the outer rotor accelerates according to formula ω. out =0.35×6000+5=2105r / min synchronous operation, the harmonic reducer reduces the output speed to 21.05r / min to drive the workpiece tray. During the grinding process, the piezoelectric ceramic fine-tuning module adjusts in real time based on the preset pressure curve: the capacitive displacement sensor monitors the actual pressure at a sampling frequency of 5kHz. When the pressure fluctuation exceeds ±1.5N, the PID controller (Kp=0.8) outputs a compensation voltage to the stacked piezoelectric actuator, which generates a ±0.8μm displacement correction to maintain the pressure stable at 75±0.5N / cm²; at the same time, the dynamic balance system continuously monitors the vibration spectrum under high-speed rotation. If the FFT analysis finds that the fundamental frequency amplitude suddenly increases to 8μm (phase angle 120°), the system immediately outputs a reverse compensation current of 4.1A to the electromagnetic unit No. 32, generating a 12N·m anti-vibration torque to cancel the disturbance.

[0040] After 120 seconds of rough grinding, the intelligent control system triggers a process switch: the pneumatic servo valve reduces the pressure from 75 N / cm² to 30 N / cm² within 20 ms, and the speed ratio k value smoothly transitions to 0.85 (target speed ω of the outer rotor) via an S-curve. out =0.85×6000+5=5105r / min). After the switch is completed, the in-situ measurement system is activated: the pneumatic lifting device raises the grinding disc by 1.5mm within 120ms, and the white light interferometer scans the workpiece surface along the spiral path (takes 12.5s), reconstructs the three-dimensional morphology and calculates the Sa value = 0.048μm (>threshold 0.05μm). Based on this, the system automatically extends the fine grinding time to 216s (original 180s×120%) and increases the speed ratio to 0.88 to enhance the edge grinding effect. At the end of the fine grinding stage, the acoustic emission sensor detects a sudden wave with an amplitude of 88dB and a duration of 65μs, which is determined to be the propagation of surface microcracks, and the system immediately stops and alarms.

[0041] Throughout the process, the fault diagnosis module operated synchronously: the current sensor detected that the third harmonic content increased from 1.2A to 1.52A (an increase of 26.7%), and the 200Hz vibration energy increased to 0.07m / s² (an increase of 22%). The system determined that the bearing was worn (code F01) and displayed a warning on the HMI interface stating "Replace NU216ECP bearing within 72 hours." Simultaneously, the diagnostic data was pushed to the MES system via the OPC UA protocol. After the equipment had accumulated 500 hours of operation, the adaptive optimization module analyzed and found that the displacement fluctuation σ... d If the value is greater than 0.18μm for three consecutive times, and the third harmonic of the current continues to exceed the standard, the grinding disc dressing program is automatically activated: the diamond dressing wheel feeds at 0.005mm / time for 120s. After dressing, the Sa value is restored to 0.009μm, and the equipment continues to enter the next workpiece processing cycle. 1. Multi-axis cooperative control X / Y / Z axis movement adopts cross-coupling control: the servo motor of the planetary roller screw (encoder resolution 17bit) receives EtherCAT bus commands. When the Z-axis piezoelectric module compensates for displacement, the X / Y axes are synchronously fine-tuned by ±0.003mm to maintain the concentricity of the grinding trajectory, ensuring that the grinding amount difference of the entire surface of the Φ50mm workpiece is <0.001mm.

[0042] 2. Dynamic balance and real-time performance guarantee The excitation coil of the electromagnetic counterweight unit has a pre-stored temperature-current compensation coefficient (-0.3% / ℃). When the temperature of the detection coil of the infrared thermal imager rises to 68℃, it automatically increases the output current to 4.1×[1+0.003×(68-25)]≈4.23A to offset the magnetic force attenuation caused by high temperature.

[0043] 3. Strategy to mitigate the impact of process switching During the process of pressure dropping from 75 N / cm² to 30 N / cm², the piezoelectric ceramic module outputs a -35V step voltage in advance (generating a -15μm displacement) to offset the flow delay of the pneumatic servo valve; when the speed ratio is switched, the external rotor driver injects a third harmonic (amplitude of the fundamental voltage 18%) to enhance torque stability in the low-speed range.

[0044] 4. Closed-loop quality feedback After the white light interferometer scans and identifies the edge collapse amount of 1.5μm, the system automatically adjusts the pressure of the piezoelectric module in different zones: the pressure in the edge area of ​​the workpiece is increased to 35N / cm² (the pressure in the center area is maintained at 30N / cm²), and the edge rotation speed ratio is increased to 0.92 through differential motion, so that the collapse amount converges to 0.3μm after the second fine grinding.

[0045] 5. Predictive maintenance execution Once the bearing wear warning is triggered, the MES system automatically schedules maintenance work orders: if the NU216ECP bearing inventory is sufficient, the maintenance robot will replace the bearing in the next shift; if the inventory is insufficient, the system will track the supplier's inventory in real time and generate an emergency purchase order, while reducing the equipment load to 80% of the rated power.

[0046] This embodiment was verified in the mass production of automotive turbocharger impellers: the single-piece processing time was reduced from the original process of 22 minutes to 16.5 minutes, the blade profile qualification rate was increased from 89.7% to 99.1%, and the equipment ran continuously for 1200 hours without any downtime.

[0047] This embodiment also provides a computer device applicable to an adjustable machining and grinding device, comprising: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to realize an adjustable machining and grinding device as proposed in the above embodiment.

[0048] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0049] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements an adjustable machining and grinding device as described in the above embodiments. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Red-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0050] In summary, this invention achieves nanoscale grinding pressure control (accuracy ±0.5N) through the synergistic innovation of a multi-degree-of-freedom adjustment mechanism and a piezoelectric ceramic fine-tuning module. Combined with the planetary roller screw double V-shaped raceway design and laser-clad WC-10Co guide rail coating (friction coefficient 0.08), it effectively eliminates the accuracy drift caused by backlash and thermal deformation in traditional screw drives, ensuring that the flatness of the carbide cutting edge is stably controlled within ±0.0005mm, improving accuracy by 4 times compared to existing equipment. At the same time, the electromagnetic dynamic balancing system (response time ≤10ms) replaces the mechanical counterweight with an active electromagnetic counterweight. By analyzing the vibration spectrum in real time using Fourier transform and outputting a reverse compensation current, it completely solves the chatter problem caused by centrifugal force deviation under high-speed conditions (10,000r / min). The measured surface roughness fluctuation of aerospace titanium alloy blades has been reduced from 35% to 2%, significantly improving the pass rate of thin-walled parts processing.

[0051] Furthermore, the deep integration of the dual-rotor permanent magnet synchronous motor and the intelligent control system breaks through the bottleneck of traditional differential motion. The inner rotor Halbach magnet array (air gap magnetic flux density 1.35T) combined with the outer rotor harmonic reduction drive achieves stepless adjustment of the speed ratio from 0.1 to 1.0 and a synchronization accuracy of ±0.03%. Combined with the S-curve acceleration and deceleration algorithm to optimize the uniformity of the grinding trajectory, the surface roughness distribution of the automotive turbocharger impeller flow channel achieves a uniformity of 94.7%, and the processing cycle is shortened by 25%. More importantly, the process database and the in-situ measurement system form a closed-loop feedback. The white light interferometer (vertical resolution 0.1nm) completes the three-dimensional morphology scan of the workpiece within 15 seconds. Based on the real-time Sa value, the fine grinding time and speed ratio are dynamically adjusted (e.g., the time is extended by 20% when Sa>0.05μm). The acoustic emission module (capture sensitivity 85dB) accurately identifies micro-crack signals, reducing the scrap rate of batch processing of irregularly shaped workpieces from 18% to 0.3%.

[0052] Ultimately, the fault diagnosis and adaptive optimization module constructs a predictive maintenance system. The random forest algorithm (50 decision trees) analyzes 7 features (such as displacement fluctuation σ) every 5 minutes. d (Exceeding 0.15μm and 3rd harmonic current limits), the system provides a 72-hour advance warning of bearing wear and automatically triggers a repair program. Combined with blockchain process logs to track the precision decay curve, the system ensures that after 1000 hours of continuous operation, the critical dimension drift is ≤±0.0005mm, the mean time between failures (MTBF) is extended from 500 hours to 3500 hours, and maintenance costs are reduced by 70%. In the manufacturing of high-value-added parts such as aerospace gears and artificial joints, the system achieves an annual comprehensive efficiency (OEE) of 89.5% for a single unit, providing a breakthrough solution for domestically produced high-end equipment in the precision manufacturing field.

[0053] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An adjustable machining and grinding device, characterized in that, include: The multi-degree-of-freedom adjustment mechanism consists of a planetary roller screw with a lead of 4mm and a repeatability of ±0.003mm and a cross roller guide, which drives the grinding disc to move along the X / Y / Z axes. The Z axis integrates a piezoelectric ceramic fine-tuning module with a resolution of 0.1μm. The dynamic balancing system consists of 32 sets of electromagnetic counterweight units distributed in a ring on the back side of the grinding disc. The centrifugal force deviation is detected in real time by Hall sensors, and the counterweight current is dynamically adjusted by a PID controller with a response time of ≤10ms. The composite drive module adopts a dual-rotor permanent magnet synchronous motor. The inner rotor is directly connected to the grinding disc to achieve stepless speed regulation from 0 to 10,000 r / min, while the outer rotor drives the workpiece tray to perform differential motion from 1 to 100 r / min through a harmonic reducer. The intelligent control system uses an embedded industrial PC to generate pressure-speed mapping curves based on a process database. It synchronously controls a pneumatic servo valve with a flow rate accuracy of ±0.05 L / min and a motor vector driver via an EtherCAT bus. The fault diagnosis module uses a high-precision current sensor with a bandwidth of 0-10 kHz and a MEMS accelerometer with a range of ±50 g to collect motor current harmonics and vibration spectra in real time. Feature quantities are extracted through 4096-point FFT analysis, and a three-step diagnostic process is performed. When the third harmonic of the current increases by more than 15% and the vibration energy in the 100-200Hz range increases by 20%, the bearing is judged to be worn, code F01; When the 800Hz vibration amplitude exceeds the baseline by 3 times and the current 5th harmonic fluctuation rate is >8%, abrasive blockage is determined, code F02; When the winding temperature rise rate is greater than 1.5℃ / s and the vibration of 200-400Hz increases by 30%, the cooling is deemed to have failed, code F03. The diagnostic results are uploaded to the MES system via the OPC UA protocol, and the local HMI displays a fault location map and handling guidelines.

2. The adjustable machining and grinding equipment as described in claim 1, characterized in that: Specific structural optimizations of the multi-degree-of-freedom adjustment mechanism: The nut of the planetary roller screw adopts a double V-shaped raceway design, in which the main raceway bears the axial load and the secondary raceway bears the radial off-center load. The rollers are arranged in the double raceway at a 45° staggered angle. The roller assembly process adopts liquid nitrogen cold assembly process. First, the nut body is cooled to -196℃, and then the pre-ground GCr15 bearing steel rollers are embedded into the raceway at an alternating angle. After the temperature rises back to room temperature, an interference fit is formed, and the fit tolerance is controlled at H6 / p5 level. The preload maintenance mechanism consists of 8 sets of butterfly springs, each with a thickness of 1.5mm, stacked in a total of 6 pieces, forming a ring array. After the initial preload is applied to 1200N by a hydraulic servo cylinder, the spring group dynamically compensates for thermal deformation during operation, with the preload fluctuation range ≤±50N. The slider substrate of the cross roller guide is made of 38CrMoAl nitrided steel. The surface is coated with a 0.3mm thick WC-10Co hard alloy coating by laser cladding technology. The cladding power is 3.5kW, the scanning speed is 8mm / s, the coating microhardness is ≥1200HV, the friction coefficient is stable in the range of 0.08 to 0.10, and the straightness of the guide rail reaches 0.002mm / m after calibration by a laser interferometer.

3. The adjustable machining and grinding equipment as described in claim 2, characterized in that: The control process of the piezoelectric ceramic fine-tuning module is further refined: The stacked piezoelectric actuator consists of 120 PZT-8 piezoelectric ceramic sheets, each 0.1 mm thick, connected in series. It generates a linear displacement output of 0–80 μm within a driving voltage range of 0–150 V. The displacement resolution is controlled at 0.0012 μm by a 16-bit DAC module. The capacitive displacement sensor has its detection plate rigidly connected to the grinding disc. It adopts a differential measurement structure, with an excitation frequency of 1 MHz and a sensitivity of 0.2 V / μm. It acquires the Z-axis displacement signal of the grinding disc in real time and converts it into a digital quantity. The control process begins with an embedded industrial PC generating a feedforward voltage command based on a preset grinding pressure curve. This command is dynamically calculated based on the displacement-voltage transfer function of the piezoelectric actuator. Subsequently, a capacitive displacement sensor feeds back the actual displacement value at a sampling frequency of 5kHz. The feedforward voltage is then corrected by a digital PID controller with a proportional coefficient Kp = 0.8, an integral time Ti = 0.05s, and a derivative time Td = 0.005s. Finally, a composite control signal is output to the high-voltage amplifier. During the transition from coarse grinding to fine grinding, when the system detects that the pressure setpoint has changed from 80 N / cm... 2 Reduced to 30 N / cm 2 The feedforward module outputs a -42V step voltage in advance to compensate for overshoot, while the PID controller resets the integral term to avoid integral saturation, achieving a smooth pressure transition with a transition time of ≤20ms and an overshoot of <5%.

4. The adjustable machining and grinding equipment as described in claim 3, characterized in that: The execution flow of the electromagnetic counterweight unit of the dynamic balancing system is specified as follows: The neodymium iron boron permanent magnets of the electromagnetic counterweight unit are fixed in the annular groove of the titanium alloy base on the back side of the grinding disc by high temperature epoxy resin. The 32 sets of electromagnetic counterweight units are distributed at equal angles. The excitation coil uses high-temperature resistant polyimide enameled wire with a diameter of 0.5mm and 120 turns, wound into a hollow cylindrical structure. The coil frame is a zirconia ceramic tube with a wall thickness of 0.8mm. When the counterweight adjustment is started, the Hall sensor embedded in the center of the coil, with a sensitivity of 1.2mV / G, collects the magnetic field vector generated by the centrifugal force in real time, and converts it into a digital signal by the 24-bit ADC module at a sampling rate of 10kHz. The signal processing unit performs a 1024-point FFT analysis on the raw data, extracting the vibration amplitude and phase angle of 1 to 3 times the fundamental frequency. When the fundamental frequency amplitude is detected to exceed 5 μm, the PID controller sets Kp = 1.5, Ki = 0.2, and Kd = 0.02, and calculates the compensation current value based on the phase angle, ranging from 0 to 5 A. This compensation current is applied to the excitation coil via the PWM drive module. The reverse electromagnetic field generated by the current couples with the permanent magnet to form a maximum anti-vibration torque of 15 N·m, which cancels out the imbalance. The system has a pre-stored static imbalance compensation curve: below 3000 r / min, linear compensation mode is used with a current gain of 0.8 A / g·mm; above 3000 r / min, it switches to quadratic function compensation with a current gain of 1.2 A / g·mm - 0.0003 A / rpm. 2 Meanwhile, the coil temperature is monitored by thermocouples and the current coefficient is dynamically corrected, with a temperature compensation rate of -0.3% / ℃.

5. The adjustable machining and grinding equipment as described in claim 4, characterized in that: The dual-rotor permanent magnet synchronous motor of the composite drive module is structurally refined and its coordinated control is implemented: The inner rotor adopts a Halbach array magnet arrangement, consisting of 16 N50SH neodymium iron boron magnets bonded sequentially to the SMC soft magnetic composite matrix with a deflection angle of 22.5° per pole, forming a unilateral enhanced magnetic field, so that the air gap magnetic flux density reaches 1.35T. The outer rotor is made of 0.1mm thick Fe-Si-Cr soft magnetic alloy sheets stacked together. 36 wedge-shaped grooves are opened on the surface of the pole shoes to suppress eddy current losses. Distributed temperature sensors are embedded in the rotor yoke. The stator winding adopts a segmented structure, with each phase consisting of 6 independent sector coils with a wire diameter of 1.2mm and 12 turns per coil. A 0.8mm cooling channel is reserved between the coils, and copper heat-conducting fins with a thickness of 0.3mm are welded to the ends of each coil. The cooling system consists of a two-stage constant temperature circulation device. The first-stage circulation cools the deionized water to 25±0.5℃ through the semiconductor refrigeration module. The second-stage circulation injects the coolant into the annular main pipe on the back of the stator through a micro turbine pump at a flow rate of 12L / min. The coolant is then distributed to each coil flow channel through 12 stainless steel capillary tubes. The outlet water temperature monitoring accuracy is ±0.3℃. The motor control adopts a dual closed-loop strategy: the inner loop is a sensorless position detection based on high-frequency signal injection with an angular resolution of 0.01°, and the outer loop achieves stepless speed regulation of the inner rotor from 0 to 10,000 r / min through field-oriented control (FOC). At the same time, the outer rotor is driven by a harmonic reducer to make differential motion of the workpiece tray. The torque coupling between the inner and outer rotors is dynamically compensated by the load observer to ensure that the speed fluctuation is ≤ ±0.05%.

6. The adjustable machining and grinding equipment as described in claim 5, characterized in that: The differential motion control process of the composite drive module is further refined as follows: The coordinated motion of the inner and outer rotors of the dual-rotor permanent magnet synchronous motor is achieved through an electronic gearbox. The specific execution process is as follows: the motion controller receives the speed ratio command set by the process parameters, generates the target speed curve of the inner rotor based on an S-curve acceleration / deceleration algorithm, and limits the acceleration to 500 rpm / s. 2 Within this range, the target speed of the outer rotor is calculated according to formula ω. out =k·ω in +b is calculated dynamically, where k is the speed ratio coefficient, and ω in ω represents the actual rotational speed of the inner rotor, where b is the base rotational speed of 1–10 r / min; ω represents the actual rotational speed of the inner rotor. in The rotational speed is detected in real time by a rotary transformer. The detection signal is converted into a digital quantity by an RDC decoding chip. The rotational speed of the outer rotor is collected by a circular grating encoder on the output shaft of the harmonic reducer, with 50,000 pulses per revolution. Speed ​​error compensation employs a feedforward-feedback composite control: the feedforward channel is based on the disturbance torque predicted by the load observer, with a sampling period of 100μs, and suppresses speed fluctuations through pre-compensation current; the feedback channel uses a dual closed-loop PID controller, with a speed loop Kp = 0.6 and an integral time of 0.1s, and a position loop Kp = 12 and an integral time of 0.05s; when the speed ratio switches from coarse grinding mode k = 0.4 to fine grinding mode k = 0.9, the system first smoothly transitions the k value with a time constant of 20ms, while dynamically adjusting the b value to compensate for inertial delay, and the external rotor acceleration is synchronously limited to 30rpm / s. 2 The control signal is transmitted to the dual-channel vector driver via the EtherCAT bus. The inner rotor driver uses space vector PWM with a switching frequency of 8kHz, while the outer rotor driver injects third harmonics to enhance low-speed torque with a carrier frequency of 4kHz.

7. An adjustable machining and grinding device as described in claim 6, characterized in that: The parameter calling and switching control of the process database are implemented in detail: The process database stores 30 sets of standard grinding formulas. Each formula includes a pressure gradient curve, four stages (rough grinding / semi-fine grinding / fine grinding / ultra-fine grinding), a speed ratio dynamic function, abrasive formula code, and processing time threshold. When parameters are called, the embedded industrial PC automatically matches the basic parameters based on the workpiece material code, and then adds correction coefficients based on the workpiece morphology scanning data. The abrasive supply system is equipped with a four-channel precision metering pump, which accurately mixes diamond micro powder (5-30%), cubic boron nitride suspension (3-15%), cerium oxide polishing slurry, and pH adjuster according to the formula code, with a mixing accuracy of ±0.5 ml / min. The parameter switching control adopts a three-level fuzzy PID strategy: the first level is during the transition from coarse grinding to semi-fine grinding, where the pressure changes from 80 N / cm². 2 →50N / cm 2 With a pressure change rate of 2.5 N / cm 2 ·s is a constraint condition. Through adaptive adjustment of proportional gain, Kp linearly decreases from 0.8 to 0.

3. In the second stage, during the transition from semi-fine grinding to fine grinding, the speed ratio k changes from 0.5 to 0.

9. A feedforward compensation algorithm is adopted, and the compensation torque is preloaded based on the speed difference. ΔT = 0.05·J·dω / dt, where J is the identified value of rotational inertia. When the third stage starts during the ultra-fine grinding stage, the in-situ measurement system is simultaneously triggered to evaluate the surface quality. If the roughness Sa > 0.02 μm, the fine grinding time is extended by 15% and the speed ratio is increased to 0.

95. If Sa ≤ 0.02 μm, the original parameters are maintained. All parameter switching processes are displayed in real time on the HMI interface, and the actual execution curve and theoretical deviation value are recorded. The pressure fluctuation is < ±1.5%, and the speed ratio error is < ±0.

01.

8. The adjustable machining and grinding equipment as described in claim 7, characterized in that: The execution flow of the in-situ measurement system is implemented in detail as follows: When the grinding stage switching signal is triggered, the pneumatic lifting device, with a cylinder diameter of Φ40mm and a stroke of 25mm, pushes the ceramic push rod with a pressure of 0.5MPa and an elastic modulus of 300GPa, and lifts the grinding disc by 1.5±0.02mm within 120ms. During the separation process, the piezoelectric ceramic fine-tuning module synchronously outputs reverse displacement to compensate for mechanical vibration. A white light interferometer, with a wavelength of 460 nm and a vertical scanning range of 500 μm, is moved to the center of the workpiece via a three-axis precision displacement stage and scans along a spiral path with a radius from 0 to R. max R max At the maximum radius, with an angular velocity of 10 rad / s, surface topography data was collected with a scan point spacing of 2.5 μm and a collection time of ≤0.8 s per frame. The topography processing unit reconstructs the three-dimensional surface based on the phase-shifting interferometry algorithm and extracts key parameters: the Sa value of the central region with an area diameter of 2 mm, the edge collapse amount of the height difference within the outermost ring width of 0.3 mm, and the total thickness change TTV; if Sa > 0.05 μm, the system automatically extends the fine grinding time by 20% and increases the rotation speed ratio by 0.05; if Sa ≤ 0.05 μm, the acoustic emission monitoring process is activated. The acoustic emission sensor, with a resonant frequency of 150kHz and a sensitivity of 85dB, is attached to the back of the workpiece tray via a magnetic base. After the signal is filtered by a 100-500kHz bandpass filter and pre-amplified by a 40dB amplifier, the characteristic parameters are analyzed in real time by the FPGA module. When a burst wave with an amplitude > 85dB and a duration > 50μs is detected, it is determined to be a crack propagation signal, and grinding is immediately stopped and an alarm is triggered. Measurement data and processing results are transmitted to the database via industrial Ethernet to update process parameters. At the same time, the HMI interface displays a three-dimensional topographic cloud map and defect location coordinates. The entire in-situ measurement cycle is controlled within 15 seconds.

9. An adjustable machining and grinding device as described in claim 8, characterized in that: The specific execution flow of the fault diagnosis module is further refined as follows: When the equipment is started, a high-precision current sensor with a bandwidth of 0 to 10 kHz collects the current signal of the three-phase winding of the dual-rotor motor in real time. Simultaneously, a MEMS accelerometer with a range of ±50g is installed on the radial and axial measuring points of the motor housing through a magnetic base. The signal processing unit acquires raw data at a sampling rate of 10kHz. After the current signal is filtered by an FIR low-pass filter with a cutoff frequency of 1.5kHz to eliminate switching noise, it is analyzed by FFT with a 4096-point Hanning window to extract the fundamental to 15th harmonic components. The vibration signal is decomposed by wavelet packet in 6 layers to extract the energy values ​​of eight sub-bands from 50 to 8000Hz. The diagnostic logic executes a three-step judgment: Step 1: Detect the third harmonic content of the current. If the increase is greater than 15% for five consecutive sampling cycles and accompanied by a 20% increase in vibration energy in the 100-200Hz range, it is judged as bearing wear, code F01, triggering a yellow warning and recommending bearing replacement within 72 hours; Step 2: Monitor the vibration spectrum. When the amplitude at the 800Hz frequency point exceeds three times the baseline value for 10 seconds, and the fifth harmonic fluctuation rate of the current is greater than 8%, it is judged as abrasive blockage, code F02, initiating the automatic cleaning program of the grinding disc, with a reverse flushing pressure of 0.8MPa for 15 seconds; Step 3: Integrate temperature data. If the stator winding temperature rise rate is greater than 1.5℃ / s and the vibration energy in the 200-400Hz range suddenly increases by 30%, it is judged as cooling failure, code F03, immediately reducing the speed to a safe value and closing the coolant valve; All diagnostic results are encapsulated into JSON format messages via the OPC UA protocol and pushed to the MES system production dashboard in real time. At the same time, the local HMI interface displays a fault location map and handling guide. The diagnostic response delay is ≤50ms, and the false alarm rate has been verified to be <0.8% after thousands of hours of testing.

10. An adjustable machining and grinding device as described in claim 9, characterized in that: The adaptive optimization module of the intelligent control system is implemented in detail as follows: During equipment operation, the embedded industrial PC collects a multi-dimensional status dataset every 5 minutes, including the displacement fluctuation σ of the piezoelectric fine-tuning module. d The electromagnetic counterweight unit current change rate di / dt, the dual rotor motor winding temperature gradient ΔT / Δt, and the in-situ measured Sa value sequence; Data preprocessing employed a sliding window filter with a window width of 10 periods and a cutoff frequency of 0.1 Hz. After noise removal, seven key features were extracted: displacement fluctuation σ. d >0.15μm is marked as characteristic F1, di / dt>2A / s is marked as F2, winding temperature rise rate>1.2℃ / s is marked as F3, Sa value increases for 3 consecutive times is marked as F4, vibration 800Hz frequency band energy increase>25% is marked as F5, current third harmonic continuously exceeds the standard is marked as F6, abrasive flow rate fluctuation>8% is marked as F7; A decision model based on the random forest algorithm was pre-trained with 50,000 samples. Real-time analysis of feature combinations was performed: when the combination [F1+F4+F6] was detected, the flatness of the grinding disc was determined to be deteriorated, and the disc dressing program was automatically activated, with the diamond dressing wheel feed rate at 0.005 mm / cycle and a dressing time of 120 s; when the combination [F2+F5+F7] appeared, the abrasive pipeline was determined to be blocked, and a three-stage pulse backflushing was executed, with pressure increasing in stages of 0.6 / 1.0 / 1.4 MPa, 5 pulses per stage; when the combination [F3+F6] lasted for 2 cycles, the bearing lubrication was determined to be ineffective, and a special grease was injected, with a dosage of 0.2 ml and a viscosity of ISO VG 32. All optimization operations are recorded in a blockchain-based process log, encrypted with SHA-256. The Sa value is re-evaluated after each optimization: if the Sa decrease is less than 10%, the expert system is activated to intervene and call the optimization parameters of similar working conditions in the case library to refine the grinding pressure or extend the time. After processing 20 workpieces, the offline calibration module automatically compares the key dimensions, flatness / cylindricity / roughness of the first and last workpieces, generates an accuracy decay curve, and reverses the weight coefficients of the decision model. The correction amount ΔW = 0.05 × |Δaccuracy|, realizing closed-loop self-evolutionary control.

Images