An adaptive vibration suppression method for diamond pick based on fuzzy control algorithm

Abstract

The present invention discloses a method for adaptive vibration suppression of diamond picks based on a fuzzy control algorithm, comprising the following steps: S1. real-time acquisition of a multi-channel signal data set during diamond pick processing; S2. acquisition of a pre-processed multi-channel signal data set; S3. integration to form a feature vector containing multi-dimensional dynamic features; S4. inputting the feature vector into a fuzzy control module, fuzzifying the feature vectors of each channel signal, and realizing fuzzy quantification of the processing state; S5. generating a control output quantity based on the fuzzy quantification result using a fuzzy inference algorithm, and converting the control output quantity into an executable process control instruction through a defuzzification algorithm; S6. feeding back the process control instruction to the processing execution mechanism. The present invention has higher sensitivity and adaptability in identifying unstable processing states, can significantly improve the accuracy of discrimination in the initial stage of vibration, and enhance the judgment basis of the control system under complex disturbances.

Description

Technical Field

[0001] The present invention relates to the field of diamond technology, in particular to a diamond pick self-adaptive vibration suppression method based on a fuzzy control algorithm. Background Art

[0002] With the development of high-end manufacturing and complex working condition cutting technology, diamond picks, as a high-hardness, high-wear-resistant precision machining tool, are widely used in high-load, high-speed metal and composite material cutting scenarios. However, in the actual processing process, since diamond picks are subjected to severe mechanical and thermal loads during cutting, they are very likely to produce nonlinear, time-varying and strong vibration phenomena. Vibration not only destroys the stability of the processing process, but also shortens the tool life and reduces cutting accuracy. In severe cases, it may even cause equipment damage and process failure.

[0003] Existing vibration suppression methods often rely on damping design of mechanical structures or traditional control strategies with fixed parameters to adjust process parameters to mitigate vibration effects. While these control methods are somewhat effective under static or regular vibration conditions, they suffer from the following drawbacks when faced with nonlinear vibrations caused by complex factors such as tool wear, workpiece material changes, temperature fluctuations, and load disturbances during diamond pick machining: First, they lack the ability to comprehensively perceive multi-source disturbance signals, failing to fully reflect the cause of vibration and the overall operating conditions of the system in real time; second, the control system is slow to respond and has fixed parameters, making it difficult to achieve rapid adaptive adjustment based on changes in the actual machining environment; and third, they fail to fully utilize the interactive characteristics of different physical signals for collaborative analysis and judgment, resulting in insufficiently targeted and precise vibration control.

[0004] In addition, some studies have attempted to apply fuzzy control algorithms to processing control to enhance the system's nonlinear processing capabilities and uncertainty adaptability. However, fuzzy control is often only performed on a single signal source, ignoring the coupling characteristics between multi-channel information during diamond cutter processing, resulting in the control strategy being unable to perform as expected in complex environments. Furthermore, the current fuzzy control methods are mostly based on static rule bases, lack self-learning capabilities, and cannot cope with real-time updates of control logic under dynamic changes in working conditions.

[0005] In summary, there is an urgent need for a new vibration suppression method that can integrate multi-channel signals and has real-time reasoning and adaptive feedback capabilities to achieve efficient and stable control of vibration problems under complex machining conditions. Summary of the Invention

[0006] One purpose of the present invention is to propose a diamond pick adaptive vibration suppression method based on a fuzzy control algorithm. The present invention has higher sensitivity and adaptability in the identification of unstable processing states, can significantly improve the accuracy of early vibration judgment, and enhance the judgment basis of the control system under complex disturbances.

[0007] According to an embodiment of the present invention, a method for adaptive vibration suppression of a diamond pick based on a fuzzy control algorithm includes the following steps:

[0008] S1. Multiple highly sensitive sensors are deployed at key locations on the diamond pick processing equipment to collect multi-channel signal data sets during the diamond pick processing process in real time;

[0009] S2. Preprocessing the multi-channel signal dataset to obtain a preprocessed multi-channel signal dataset;

[0010] S3. Extract vibration characteristic parameters, temperature characteristic parameters, and load characteristic parameters from the preprocessed multi-channel signal dataset and integrate them to form a feature vector containing multi-dimensional dynamic characteristics;

[0011] S4. The feature vector is input to the fuzzy control module. The fuzzy control module performs fuzzy processing on the feature vectors of each channel signal according to the preset multi-channel fuzzy control rule base and membership function to achieve fuzzy quantification of the processing state;

[0012] S5. Based on the fuzzy quantization results, a fuzzy inference algorithm is used to generate a control output, and a defuzzification algorithm is used to convert the control output into an executable process control instruction;

[0013] S6. Feedback process control instructions to the machining actuator to adjust key process parameters in real time, including tool speed, feed rate, and cooling parameters, to achieve active suppression of vibration status, temperature, and load condition parameters.

[0014] Optionally, the S1 includes the following steps:

[0015] S11. Arrange multiple sensor arrays at the spindle end of the diamond pick processing equipment, the tool fixing structure and the workpiece support area, the sensor array includes a vibration sensor, a temperature sensor and a load sensor;

[0016] S12. During the diamond pick cutting process, the original data stream of the multi-channel signal is acquired at fixed time intervals to construct a multi-channel signal data set. , each data in the multi-channel signal dataset This corresponds to the physical state of different sensor collection positions during diamond pick processing, and is used to reflect the real-time dynamic working conditions of the processing:

[0017]

[0018] in, Indicates the Multi-channel signal data of sampling points, is the timestamp of the sampling point; Indicates the A vibration sensor at time The collected vibration signal, Indicates the A temperature sensor at time The collected temperature signal, Indicates the The load sensors are The collected load signal, is the total number of sampling points.

[0019] Optionally, the S2 includes the following steps:

[0020] S21. Multi-channel signal dataset Each type of sensor signal in the system is band-pass filtered separately. The frequency range retained for the vibration signal is the frequency range between the minimum and maximum effective frequencies of the vibration signal. The temperature signal and the load signal are also set with corresponding minimum and maximum effective frequencies. The noise and interference components in the non-working frequency band of each channel signal are filtered out to obtain a filtered multi-channel signal data set.

[0021] S22. De-noise each sensor signal after bandpass filtering. For each sampling point, take the average of multiple consecutive sampling points within a time window of a set length to smooth local fluctuations in each channel signal, thereby obtaining a multi-channel signal dataset consisting of denoised vibration, temperature, and load signals.

[0022] S23. Normalize the denoised multi-channel signal dataset separately, and recombine the normalized vibration signal, temperature signal, and load signal according to the corresponding timestamps to form a preprocessed multi-channel signal dataset. .

[0023] Optionally, S3 includes the following steps:

[0024] S31. Extract features based on the preprocessed multi-channel signal data set to obtain the vibration feature parameter set ,The vibration characteristic parameter set includes the root mean square value representing the ,energy intensity of the vibration signal, the crest factor representing the ratio between the ,maximum amplitude and the effective value of the signal, the main frequency ,component representing the frequency where the energy of the signal is concentrated in ,the frequency domain, and the kurtosis and skewness used to measure the ,non-Gaussianity and symmetry of the signal;

[0025] S32. Perform feature extraction based on the preprocessed multi-channel signal data set to extract the temperature feature parameter set ,The temperature characteristic parameter set includes the instantaneous temperature rise rate representing the ,temperature change rate between two adjacent time points, the average temperature value of ,the sliding window to reflect the local stable thermal state, and the ,temperature fluctuation amplitude to characterize the local thermal heterogeneity;

[0026] S33. Perform feature extraction based on the preprocessed multi-channel signal data set to extract the load feature parameter set ,The set of load characteristic parameters that represents the average load level during machining includes the average cutting force, the cutting force fluctuation rate that represents the degree of load change per unit time, and the load peak duration that represents the duration of maintenance in the overload state;

[0027] S34. Each time point The corresponding vibration characteristic parameter set , temperature characteristic parameter set and load characteristic parameter sets Fusion and integration to form a feature vector containing multi-dimensional dynamic features .

[0028] Optionally, the S4 includes the following steps:

[0029] S41. The feature vector of the multi-dimensional dynamic feature Input to the fuzzy control module as the joint input of fuzzy control;

[0030] S42. Establish fuzzy linguistic variables for each feature input variable in the feature vector of the multidimensional dynamic feature, set the fuzzy linguistic value level to low, medium, and high, and use membership functions to fuzzify the feature input variables to form a fuzzy input set;

[0031] S43. Construct a fuzzy control rule base. The fuzzy control rules in the fuzzy control rule base are established based on actual processing experience and historical data analysis. The input conditions are composed of fuzzy levels of three channels: vibration, temperature, and load. The output control level reflects the stability of the current processing state:

[0032] If the RMS value of the vibration signal is high, the instantaneous temperature rise rate is medium, and the cutting force fluctuation rate of the load signal is high, the machining state is unstable;

[0033] If the main frequency component of the vibration signal is medium, the average temperature is low, and the average value of the load signal is medium, the machining state is stable;

[0034] If the kurtosis of the vibration signal is low, the temperature fluctuation amplitude is high, and the peak duration of the load signal is medium, the machining status is average;

[0035] S44. Based on the membership function and fuzzy control rule base, fuzzy reasoning is performed on the characteristic input variables to generate fuzzy output quantities. The fuzzy level of the processing state is set to three levels: stable, general, and unstable.

[0036] Optionally, the S5 includes the following steps:

[0037] S51. Establish the control output fuzzy set according to the fuzzy level of the processing state , define the tool speed adjustment amount according to the machining status level stable, general, and unstable , Feed speed adjustment Adjustment of cooling parameters The fuzzy output levels of are decrease, unchanged, and increase;

[0038] S52. Based on the fuzzy inference results of the stable, general, and unstable machining state levels, control output membership functions are constructed respectively, where the value range of the control output membership function represents the adjustment amount of the tool speed, feed rate, and cooling parameter respectively;

[0039] S53. Fuzzy set of control output Defuzzification is performed and the precise value of each control output is determined based on the improved center of gravity defuzzification algorithm with dynamic weight correction. :

[0040]

[0041] in, Indicates the first adjustment value corresponding to the tool speed adjustment value, feed speed adjustment value, and cooling parameter adjustment value. The control value is The membership degree of the control value corresponding to the processing state level is stable, general or unstable. is the number of elements in the output fuzzy set, is the dynamic weight coefficient, Combined with the nonlinear degree of the actual vibration characteristics of diamond pick processing, the weight coefficient is dynamically adjusted according to the kurtosis and peak factor of the real-time monitored vibration signal, so that the weight coefficient of the control value required for the output fuzzy set close to the actual vibration state is increased, and the weight coefficient of the control value required for the deviated vibration state is reduced;

[0042] S54. The precise value of each control output Converted into a specific process control instruction set , including tool speed control instructions adjusted according to the machining status level , used to suppress vibration amplitude in real time; feed speed control instruction adjusted according to processing status level , used to adapt to load changes; cooling parameter control instructions adjusted according to the processing status level , which is used to regulate the processing temperature state and realize real-time active suppression of vibration during the diamond pick processing process.

[0043] The beneficial effects of the present invention are:

[0044] (1) The present invention introduces three types of high-frequency sampling signals, namely vibration, temperature and load, to construct a unified multi-channel signal data set, and designs a multi-dimensional dynamic feature extraction method. In the fuzzy control input stage, the vibration energy, temperature rise rate and load fluctuation trend parameters are simultaneously considered, and the fusion constructs a feature vector for fuzzy reasoning. Compared with the existing fuzzy control method that only relies on a single vibration channel, the present invention has higher sensitivity and adaptability in the identification of unstable processing states, can significantly improve the judgment accuracy of the initial vibration, and enhance the judgment basis of the control system under complex disturbances.

[0045] (2) The present invention constructs a hierarchical fuzzy rule base covering stable, general and unstable processing states. The rule input is based on the composite characteristics of vibration root mean square value, temperature fluctuation amplitude and load continuous peak value, and the output is directly mapped to the actual process response variables such as tool speed, feed speed and cooling flow adjustment direction. The fuzzy reasoning process dynamically adjusts the weight through the membership function, thereby achieving accurate mapping of state changes to control instructions, opening up the logical chain from signal recognition to process control, and improving the interpretation transparency and consistency of the control response of the fuzzy control system.

[0046] (3) The present invention uses the center of gravity method to defuzzify the fuzzy control output in the control output stage, accurately calculates the adjustment amount of tool speed, feed speed and cooling parameters, and constructs a numerical mapping model between the processing state level and the actuator instruction to form a stable control closed loop, which solves the problems of unclear control force and low feedback accuracy at the execution end of the traditional fuzzy control system, improves the system's response sensitivity and dynamic control accuracy, and effectively ensures the processing stability and anti-interference ability of diamond picks under high load environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0048] Figure 1 This is a flow chart of a diamond pick adaptive vibration suppression method based on fuzzy control algorithm proposed by the present invention. DETAILED DESCRIPTION

[0049] The present invention will now be described in further detail with reference to the accompanying drawings, which are simplified schematic diagrams that illustrate the basic structure of the present invention in a schematic manner.

[0050] refer to Figure 1 A method for adaptive vibration suppression of diamond picks based on fuzzy control algorithm comprises the following steps:

[0051] S1. Multiple highly sensitive sensors are deployed at key locations on the diamond pick processing equipment to collect multi-channel signal data sets during the diamond pick processing process in real time;

[0052] S2. Preprocessing the multi-channel signal dataset to obtain a preprocessed multi-channel signal dataset;

[0053] S3. Extract vibration characteristic parameters, temperature characteristic parameters, and load characteristic parameters from the preprocessed multi-channel signal dataset and integrate them to form a feature vector containing multi-dimensional dynamic characteristics;

[0054] S4. The feature vector is input to the fuzzy control module. The fuzzy control module performs fuzzy processing on the feature vectors of each channel signal according to the preset multi-channel fuzzy control rule base and membership function to achieve fuzzy quantification of the processing state;

[0055] S5. Based on the fuzzy quantization results, a fuzzy inference algorithm is used to generate a control output, and a defuzzification algorithm is used to convert the control output into an executable process control instruction;

[0056] S6. Feedback process control instructions to the machining actuator to adjust key process parameters in real time, including tool speed, feed rate, and cooling parameters, to achieve active suppression of vibration status, temperature, and load condition parameters.

[0057] In this embodiment, S1 includes the following steps:

[0058] S11. Multiple sensor arrays are arranged at the spindle end, tool fixing structure and workpiece support area of ​​the diamond pick processing equipment. The sensor array includes vibration sensors, temperature sensors and load sensors;

[0059] S12. During the diamond pick cutting process, the original data stream of the multi-channel signal is acquired at fixed time intervals to construct a multi-channel signal data set. , each data in the multi-channel signal dataset This corresponds to the physical state of different sensor collection positions during diamond pick processing, and is used to reflect the real-time dynamic working conditions of the processing:

[0060]

[0061] in, Indicates the Multi-channel signal data of sampling points, is the timestamp of the sampling point; Indicates the A vibration sensor at time The collected vibration signal, Indicates the A temperature sensor at time The collected temperature signal, Indicates the The load sensors are The collected load signal, is the total number of sampling points.

[0062] This embodiment introduces three types of high-frequency sampling signals, namely vibration, temperature and load, to construct a unified multi-channel signal data set, and designs a multi-dimensional dynamic feature extraction method. In the fuzzy control input stage, the vibration energy, temperature rise change rate and load fluctuation trend parameters are simultaneously considered, and a fusion feature vector is constructed for fuzzy reasoning. Compared with the existing fuzzy control method that only relies on a single vibration channel, the present invention has higher sensitivity and adaptability in the identification of unstable processing states, can significantly improve the judgment accuracy in the early stage of vibration, and enhance the judgment basis of the control system under complex disturbances.

[0063] In this embodiment, S2 includes the following steps:

[0064] S21. Multi-channel signal dataset Each type of sensor signal in the system is band-pass filtered separately. The frequency range retained for the vibration signal is the frequency range between the minimum and maximum effective frequencies of the vibration signal. The temperature signal and the load signal are also set with corresponding minimum and maximum effective frequencies. The noise and interference components in the non-working frequency band of each channel signal are filtered out to obtain a filtered multi-channel signal data set.

[0065] S22. De-noise each sensor signal after bandpass filtering. For each sampling point, take the average of multiple consecutive sampling points within a time window of a set length to smooth local fluctuations in each channel signal, thereby obtaining a multi-channel signal dataset consisting of denoised vibration, temperature, and load signals.

[0066] S23. Perform linear range normalization on the denoised multi-channel signal dataset. All signals must be independently normalized within the channel, and cross-channel normalization is prohibited. The normalized vibration signal, temperature signal, and load signal are recombined according to the corresponding timestamps to form a preprocessed multi-channel signal dataset. .

[0067] This embodiment significantly improves data quality and reliability by filtering, denoising, and normalizing multi-channel signal data sets. By performing bandpass filtering and sliding denoising on the vibration, temperature, and load signals, respectively, it effectively removes abnormal fluctuations caused by processing environment interference, mechanical resonance, or electrical noise, significantly improving the stability and accuracy of the subsequent fuzzy control system input data. The system's sensitivity to subtle vibration changes is enhanced. The filtered vibration signal retains high-frequency details related to tool wear, workpiece hard point contact, and other factors, facilitating the timely identification of small but significant vibration changes, enabling more precise vibration response control, and achieving uniform multi-channel signal normalization. By normalizing data of different dimensions and amplitudes, the problem of inconsistent scales for the three types of vibration, temperature, and load signals is effectively resolved. This provides an equally weighted and equivalent input basis for the fuzzy control module, improving the accuracy and generalization of fuzzy reasoning.

[0068] In this embodiment, S3 includes the following steps:

[0069] S31. Use the root mean square method to extract features from the preprocessed multi-channel signal data set to obtain the vibration feature parameter set ,The vibration characteristic parameter set includes the root mean square value representing the ,energy intensity of the vibration signal, the crest factor representing the ratio between the ,maximum amplitude and the effective value of the signal, the main frequency ,component representing the frequency where the energy of the signal is concentrated in ,the frequency domain, and the kurtosis and skewness used to measure the ,non-Gaussianity and symmetry of the signal;

[0070] S32. Perform feature extraction based on the preprocessed multi-channel signal data set to extract the temperature feature parameter set ,The temperature characteristic parameter set includes the instantaneous temperature rise rate representing the ,temperature change rate between two adjacent time points, the average temperature value of ,the sliding window to reflect the local stable thermal state, and the ,temperature fluctuation amplitude to characterize the local thermal heterogeneity;

[0071] S33. Perform feature extraction based on the preprocessed multi-channel signal data set to extract the load feature parameter set ,The set of load characteristic parameters that represents the average load level during machining includes the average cutting force, the cutting force fluctuation rate that represents the degree of load change per unit time, and the load peak duration that represents the duration of maintenance in the overload state;

[0072] S34. Each time point The corresponding vibration characteristic parameter set , temperature characteristic parameter set and load characteristic parameter sets Fusion and integration to form a feature vector containing multi-dimensional dynamic features .

[0073] This implementation enhances the multidimensional characterization of vibration states by extracting multi-channel feature parameters and constructing feature vectors based on preprocessed data. By extracting multiple statistical and frequency-domain indicators, including the RMS value, peak factor, dominant frequency, kurtosis, and skewness, it comprehensively characterizes the energy level, non-Gaussianity, and periodicity of the vibration signal, enhancing the ability to perceive complex vibration states. Temperature and load characteristics are introduced to achieve coupled state expression. Vibration during machining is often affected by the coupling of temperature rise and sudden load fluctuations. The introduction of parameters such as temperature rise rate, temperature fluctuation, and cutting force fluctuation enables the control system to more accurately understand the physical mechanisms behind vibration, expanding from "single control" to "coupled control." A unified time-series dynamic feature vector is constructed. By integrating the feature parameters of all channels at each sampling moment, a dynamic feature vector is formed. This provides high-dimensional, time-series continuous input to the subsequent fuzzy control module, imbuing the fuzzy control decision with time continuity and dynamic sensitivity, helping to improve the overall system's response speed and adjustment accuracy.

[0074] In this embodiment, S4 includes the following steps:

[0075] S41. The feature vector of the multi-dimensional dynamic feature Input to the fuzzy control module as the joint input of fuzzy control;

[0076] S42. Establish fuzzy linguistic variables for each feature input variable in the feature vector of the multidimensional dynamic feature, set the fuzzy linguistic value level to low, medium, and high, and use membership functions to fuzzify the feature input variables to form a fuzzy input set;

[0077] The fuzzy linguistic variables are:

[0078] Vibration characteristic variable (vibration root mean square value), fuzzy language variable name: vibration intensity, fuzzy language value level: {low, medium, high}, typical explanation: low: small vibration energy, stable processing process, medium: slight signs of instability, high: severe vibration, there is a risk of instability

[0079] Temperature characteristic variable (instantaneous temperature rise rate), fuzzy language variable name: temperature rise rate, fuzzy language value level: {slow, medium, rapid}, typical explanation: Slow: temperature changes smoothly, thermal stability is good, medium: local heat accumulation may occur, rapid: thermal effect between tool and workpiece is obvious, cooling control is required

[0080] Load characteristic variable (load fluctuation rate), fuzzy language variable name: load variation amplitude, fuzzy language value level: {small, medium, large}, typical explanation: small: load is stable and processing load is controllable, medium: load fluctuates and enters the control boundary, large: cutting force is unstable and has a tendency to get out of control.

[0081] A membership function is a mathematical tool used in fuzzy control systems to map precise numerical inputs (such as vibration intensity, temperature change rate, or load fluctuation) to fuzzy linguistic variables (low, medium, or high). Specifically, a membership function defines the degree to which an input variable belongs to a fuzzy set (i.e., its membership degree w). In this implementation, a triangular membership function is used.

[0082] S43. Construct a fuzzy control rule base. The fuzzy control rules in the fuzzy control rule base are established based on actual processing experience and historical data analysis. The input conditions are composed of fuzzy levels of three channels: vibration, temperature, and load. The output control level reflects the stability of the current processing state:

[0083] If the RMS value of the vibration signal is high, the instantaneous temperature rise rate is medium, and the cutting force fluctuation rate of the load signal is high, the machining state is unstable;

[0084] If the main frequency component of the vibration signal is medium, the average temperature is low, and the average value of the load signal is medium, the machining state is stable;

[0085] If the kurtosis of the vibration signal is low, the temperature fluctuation amplitude is high, and the peak duration of the load signal is medium, the machining status is average;

[0086] S44. Based on the membership function and fuzzy control rule base, fuzzy reasoning is performed on the characteristic input variables to generate fuzzy output quantities. The fuzzy level of the processing state is set to three levels: stable, general, and unstable.

[0087] In this embodiment, fuzzy reasoning:

[0088] The current moment The language values ​​corresponding to each fuzzy input variable are matched with the condition items in the rule base to identify the rules that meet the premise conditions.

[0089] In this embodiment, the current input variables after fuzzification are: vibration root mean square = high (membership: 0.8), temperature rise rate = medium (membership: 0.6), load fluctuation rate = high (membership: 0.9);

[0090] Then and rule: If "high" "medium" "high", then the state is "unstable" - activated

[0091] Calculate rule activation strength (inference strength)

[0092] Use the "minimum membership method" to perform conjunction logic processing (AND relationship): , indicating that this rule is activated with a strength of 0.6.

[0093] The inference outputs a fuzzy set, which assigns the above activation intensity to the output item of the rule, that is, the fuzzy set "processing state = unstable", and its membership function will be truncated or weighted on this basis.

[0094] If multiple rules are activated at the same time, the outputs of all rules are synthesized using "maximization" to finally obtain a joint output of the processing state fuzzy set:

[0095] All the rules are combined with the maximum membership of the inference results of "stable", "general" and "unstable", and finally a comprehensive output fuzzy set is formed. , which is used for subsequent defuzzification.

[0096] This implementation method constructs a hierarchical fuzzy rule library covering stable, general and unstable processing states. The rule input is based on complex characteristics such as the root mean square value of vibration, temperature fluctuation amplitude and continuous load peak, and the output is directly mapped to the actual process response variables such as tool speed, feed speed and adjustment direction of cooling flow. The fuzzy inference process dynamically adjusts the weight through the membership function, thereby achieving accurate mapping of state changes to control instructions, opening up the logical chain from signal recognition to process control, and improving the interpretation transparency and consistency of the fuzzy control system.

[0097] In this embodiment, S5 includes the following steps:

[0098] S51. Establish the control output fuzzy set according to the fuzzy level of the processing state , define the tool speed adjustment amount according to the machining status level stable, general, and unstable , Feed speed adjustment Adjustment of cooling parameters The fuzzy output levels of are decrease, unchanged, and increase;

[0099] S52. Based on the fuzzy inference results of the stable, general, and unstable machining state levels, control output membership functions are constructed respectively, where the value range of the control output membership function represents the adjustment amount of the tool speed, feed rate, and cooling parameter respectively;

[0100] S53. Fuzzy set of control output Defuzzification is performed and the precise value of each control output is determined based on the improved center of gravity defuzzification algorithm with dynamic weight correction. :

[0101]

[0102] in, Indicates the first adjustment value corresponding to the tool speed adjustment value, feed speed adjustment value, and cooling parameter adjustment value. The control value is The membership degree of the control value corresponding to the processing state level is stable, general or unstable. is the number of elements in the output fuzzy set, is the dynamic weight coefficient, Combined with the nonlinear degree of the actual vibration characteristics of diamond pick processing, the weight coefficient is dynamically adjusted according to the kurtosis and peak factor of the real-time monitored vibration signal, so that the weight coefficient of the control value required for the output fuzzy set close to the actual vibration state is increased, and the weight coefficient of the control value required for the deviated vibration state is reduced;

[0103] The improved center-of-gravity defuzzification algorithm based on dynamic weight correction proposed by the formula can significantly improve the accuracy and pertinence of the vibration suppression control output during diamond pick machining. Compared with the traditional center-of-gravity method, the algorithm can adaptively adjust the weight coefficient based on the differences in the vibration signal characteristics monitored in real time, accurately reflecting the changes in actual machining conditions, and effectively avoiding the defects of the traditional method of indiscriminately averaging the various fuzzy control quantities. Through this method, the key process control quantities such as tool speed, feed rate, and cooling parameters can be accurately adjusted in real time, effectively improving the active suppression of machining vibration, significantly extending tool life, and improving machining accuracy.

[0104] S54. The precise value of each control output Converted into a specific process control instruction set , including tool speed control instructions adjusted according to the machining status level , used to suppress vibration amplitude in real time; feed speed control instruction adjusted according to processing status level , used to adapt to load changes; cooling parameter control instructions adjusted according to the processing status level , which is used to regulate the processing temperature state and realize real-time active suppression of vibration during the diamond pick processing process.

[0105] This implementation uses the center of gravity method to defuzzify the fuzzy control output in the control output stage, accurately calculates the adjustment amounts of tool speed, feed speed and cooling parameters, and constructs a numerical mapping model between the processing state level and the actuator instructions to form a stable control closed loop. It solves the problems of unclear control force and low feedback accuracy at the execution end of traditional fuzzy control systems, improves the system's response sensitivity and dynamic control accuracy, and effectively ensures the processing stability and anti-interference ability of diamond picks under high-load environments.

[0106] Example 1:

[0107] At 9:32 AM on November 18, 2024, a high-precision equipment manufacturer in Wuxi, Jiangsu Province, activated the diamond pick adaptive vibration suppression system, which utilizes the present invention, while machining a batch of critical aviation structural parts. The task required six hours of continuous precision cutting of a batch of high-strength titanium alloy components, with machining errors controlled within ±20μm.

[0108] During this task, the system operated stably during the initial cutting phase (9:32:00-9:32:10). At 9:32:11, the system detected an abnormal signal from the vibration sensor (S_v_04): the peak acceleration reached 1.62g, significantly higher than the historical average of 0.9g. The sampled data is shown below:

[0109] ;

[0110] The system makes fuzzy control decisions based on the vibration root mean square value (RMS = 0.45g), temperature rise rate (2.8°C / s), and cutting force fluctuation rate (14.2%) at that moment. According to the preset fuzzy rule base, the following conditions are met:

[0111] If the RMS is high, the temperature rise rate is medium, and the fluctuation rate is high, the state is "unstable".

[0112] The system determines that the current processing state is "unstable" and generates a fuzzy output set :

[0113] The tool speed adjustment level is "reduced";

[0114] The feed speed adjustment level is "reduced";

[0115] Cooling parameter adjustment level is "Increase".

[0116] The system then calls the center of gravity method for defuzzification, generating the following precise control output:

[0117] ;

[0118] At 9:32:13, this set of control instructions was sent to the execution controller of the machining center through the system bus, completing the dynamic adjustment of process parameters in real time, avoiding the risk of tool jumping or local deformation caused by excessive load and vibration accumulation.

[0119] At 9:36:47, the system again detected an abnormal fluctuation from the temperature sensor (No.: S_t_02). The average temperature within the sliding window rose rapidly from 78.6°C to 89.4°C, with a temperature rise rate exceeding 3.3°C / s. At the same time, the peak load feedback from the load sensor (No.: S_f_01) lasted for more than 4.2 seconds, far exceeding the historical average of 1.7 seconds.

[0120] The fuzzy reasoning result is again evaluated as "unstable", and the system immediately generates dynamic instructions again. The control output of this round is as follows:

[0121] ;

[0122] At the same time, the system records the complete abnormal segment data in the background and automatically generates a diagnostic report on the control terminal, prompting that the processing process has entered the "heat-load-vibration coupling abnormal section" and recommending that the cutting path be optimized first in subsequent batches.

[0123] In the control group of the same batch that did not use the present invention, the following data comparison differences were found for the same workpieces recorded during the same time period:

[0124]

[0125] According to process records, the method of the present invention identified 47 abnormal vibration, heat, and load combinations during the entire 6-hour machining task, completing all of them with closed-loop response control. The average initial control response took 1.7 seconds, significantly shorter than the 5.4-second average response delay of traditional methods. Ultimately, the task resulted in the processing and delivery of a batch of 98 aviation parts. After inspection, all dimensional errors were controlled within a range of ±17μm, exceeding the company's standard of ±20μm. Tool wear was reduced by 28%, and the frequency of tool changes was reduced by nearly one-third.

[0126] In summary, this embodiment truly demonstrates the online application process, vibration control closed-loop response, fuzzy reasoning decision-making process and control instruction implementation effect of the method of the present invention in a high-precision diamond cutting scenario, and verifies the multiple advantages of the present invention in terms of robustness, response speed and control accuracy in a dynamic nonlinear machining environment.

[0127] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A diamond pick adaptive vibration suppression method based on fuzzy control algorithm, characterized in that: The steps include: S1. Multiple highly sensitive sensors are deployed at key locations on the diamond pick processing equipment to collect multi-channel signal data sets during the diamond pick processing process in real time; S2. Preprocessing the multi-channel signal dataset to obtain a preprocessed multi-channel signal dataset; S3. Extract vibration characteristic parameters, temperature characteristic parameters, and load characteristic parameters from the preprocessed multi-channel signal dataset and integrate them to form a feature vector containing multi-dimensional dynamic characteristics; S4. The feature vector is input to the fuzzy control module. The fuzzy control module performs fuzzy processing on the feature vectors of each channel signal according to the preset multi-channel fuzzy control rule base and membership function to achieve fuzzy quantification of the processing state; S5. Based on the fuzzy quantization results, a fuzzy inference algorithm is used to generate a control output, and a defuzzification algorithm is used to convert the control output into an executable process control instruction; S6. Feedback process control instructions to the machining actuator to adjust key process parameters in real time, including tool speed, feed rate, and cooling parameters, to achieve active suppression of vibration, temperature, and load parameters. Said S1 comprises the following steps: S11. Arrange multiple sensor arrays at the spindle end of the diamond pick processing equipment, the tool fixing structure and the workpiece support area, the sensor array includes a vibration sensor, a temperature sensor and a load sensor; S12. During the diamond pick cutting process, the original data stream of the multi-channel signal is acquired at fixed time intervals to construct a multi-channel signal data set. , each data in the multi-channel signal dataset This corresponds to the physical state of different sensor collection positions during diamond pick processing, and is used to reflect the real-time dynamic working conditions of the processing: in, Indicates the Multi-channel signal data of sampling points, is the timestamp of the sampling point; Indicates the A vibration sensor at time The collected vibration signal, Indicates the A temperature sensor at time The collected temperature signal, Indicates the The load sensors are The collected load signal, is the total number of sampling points; The S2 comprises the following steps: S21. Multi-channel signal dataset Each type of sensor signal in the system is band-pass filtered separately. The frequency range retained for the vibration signal is the frequency range between the minimum and maximum effective frequencies of the vibration signal. The temperature signal and the load signal are also set with corresponding minimum and maximum effective frequencies. The noise and interference components in the non-working frequency band of each channel signal are filtered out to obtain a filtered multi-channel signal data set. S22. De-noise each sensor signal after bandpass filtering. For each sampling point, take the average of multiple consecutive sampling points within a time window of a set length to smooth local fluctuations in each channel signal, thereby obtaining a multi-channel signal dataset consisting of denoised vibration, temperature, and load signals. S23. Normalize the denoised multi-channel signal dataset separately, and recombine the normalized vibration signal, temperature signal, and load signal according to the corresponding timestamps to form a preprocessed multi-channel signal dataset. ; The S3 includes the following steps: S31. Extract features based on the preprocessed multi-channel signal data set to obtain the vibration feature parameter set ,The vibration characteristic parameter set includes the root mean square value representing the ,energy intensity of the vibration signal, the crest factor representing the ratio between the ,maximum amplitude and the effective value of the signal, the main frequency ,component representing the frequency where the energy of the signal is concentrated in ,the frequency domain, and the kurtosis and skewness used to measure the ,non-Gaussianity and symmetry of the signal; S32. Perform feature extraction based on the preprocessed multi-channel signal data set to extract the temperature feature parameter set ,The temperature characteristic parameter set includes the instantaneous temperature rise rate representing the ,temperature change rate between two adjacent time points, the average temperature value of ,the sliding window to reflect the local stable thermal state, and the ,temperature fluctuation amplitude to characterize the local thermal heterogeneity; S33. Perform feature extraction based on the preprocessed multi-channel signal data set to extract the load feature parameter set ,The set of load characteristic parameters that represents the average load level during machining includes the average cutting force, the cutting force fluctuation rate that represents the degree of load change per unit time, and the load peak duration that represents the duration of maintenance in the overload state; S34. Each time point The corresponding vibration characteristic parameter set , temperature characteristic parameter set and load characteristic parameter sets Fusion and integration to form a feature vector containing multi-dimensional dynamic features ; The S4 comprises the following steps: S41. The feature vector of the multi-dimensional dynamic feature Input to the fuzzy control module as the joint input of fuzzy control; S42. Establish fuzzy linguistic variables for each feature input variable in the feature vector of the multidimensional dynamic feature, set the fuzzy linguistic value level to low, medium, and high, and use membership functions to fuzzify the feature input variables to form a fuzzy input set; S43. Construct a fuzzy control rule base. The fuzzy control rules in the fuzzy control rule base are established based on actual processing experience and historical data analysis. The input conditions are composed of fuzzy levels of three channels: vibration, temperature, and load. The output control level reflects the stability of the current processing state. S44. Based on the membership function and fuzzy control rule base, fuzzy inference is performed on the characteristic input variables to generate fuzzy output quantities. The fuzzy level of the processing state is set to three levels: stable, general, and unstable; The S5 comprises the following steps: S51. Establish the control output fuzzy set according to the fuzzy level of the processing state , define the tool speed adjustment amount according to the machining status level stable, general, and unstable , Feed speed adjustment Adjustment of cooling parameters The fuzzy output levels of are decrease, unchanged, and increase; S52. Based on the fuzzy inference results of the stable, general, and unstable machining state levels, control output membership functions are constructed respectively, where the value range of the control output membership function represents the adjustment amount of the tool speed, feed rate, and cooling parameter respectively; S53. Fuzzy set of control output Defuzzification is performed and the precise value of each control output is determined based on the improved center of gravity defuzzification algorithm with dynamic weight correction. : in, Indicates the first adjustment value corresponding to the tool speed adjustment value, feed speed adjustment value, and cooling parameter adjustment value. The control value is The membership degree of the control value corresponding to the processing state level is stable, general or unstable. is the number of elements in the output fuzzy set, is the dynamic weight coefficient, Combined with the nonlinear degree of the actual vibration characteristics of diamond pick processing, the weight coefficient is dynamically adjusted according to the kurtosis and peak factor of the real-time monitored vibration signal, so that the weight coefficient of the control value required for the output fuzzy set close to the actual vibration state is increased, and the weight coefficient of the control value required for the deviated vibration state is reduced; S54. The precise value of each control output Converted into a specific process control instruction set , including tool speed control instructions adjusted according to the machining status level , used to suppress vibration amplitude in real time; feed speed control instruction adjusted according to processing status level , used to adapt to load changes; cooling parameter control instructions adjusted according to the processing status level , which is used to regulate the processing temperature state and realize real-time active suppression of vibration during the diamond pick processing process.

2. The method for adaptive vibration suppression of diamond picks based on fuzzy control algorithm according to claim 1, characterized in that: The three levels of stable, general and unstable are constructed as follows: If the RMS value of the vibration signal is high, the instantaneous temperature rise rate is medium, and the cutting force fluctuation rate of the load signal is high, the machining state is unstable; If the main frequency component of the vibration signal is medium, the average temperature is low, and the average value of the load signal is medium, the machining state is stable; If the kurtosis of the vibration signal is low, the temperature fluctuation amplitude is high, and the peak duration of the load signal is medium, the machining status is average.

Images