Alloy saw blade production and processing automatic control method and system

Abstract

The application relates to the technical field of mechanical processing control, and discloses an alloy saw blade production and processing automatic control method and system.The method comprises the following steps: determining abnormal frequency distribution by acquiring an original vibration signal and converting the original vibration signal into a frequency domain spectrum; extracting a feature vector and determining an abnormal separation vector when an energy proportion exceeds a limit; calculating the deviation of the feature vector from a normal mode by using a support vector machine, combining a fault mechanism topology graph to determine a coupling influence index; adjusting the rotating speed and the feed amount according to a leading abnormal source and a machine tool stiffness characteristic to determine an adjustment scheme and update equipment control instructions; and adopting wavelet transform to analyze signal multi-scale components, and realizing minimum coupling influence control of the whole intermittent cutting process through an iterative optimization cycle condition. The method can solve the problem of low early abnormal identification precision in the prior art.

Description

Technical Field

[0001] This invention relates to the field of machining control technology, and in particular to an automated control method and system for the production and processing of alloy saw blades. Background Technology

[0002] Currently, alloy saw blades are a key cutting tool in modern manufacturing, and their production and processing directly determine product quality and equipment lifespan. Under the demands of high-precision, high-efficiency mass production, industrial control systems must ensure that saw blade processing equipment operates continuously and stably for extended periods under high-speed, heavy-load, intermittent cutting conditions. This places extremely high demands on the reliability of automated control and the level of intelligent monitoring.

[0003] In existing technologies, the status monitoring of saw blade processing equipment mainly relies on manual experience or simple preset threshold alarm methods. This approach collects vibration amplitude values ​​during processing using sensors and triggers a shutdown or alarm when the value exceeds a fixed threshold. However, carbide saw blade processing is a complex intermittent cutting process, and the resulting vibration signals are highly time-varying and coupled. The signals contain both the fundamental frequency and its harmonic components reflecting the normal cutting state, as well as non-stationary transient characteristic frequencies caused by saw tooth wear, chipping, or changes in vibration modes. Existing technologies lack the ability to quantify and analyze the deep characteristics of complex vibration signals, making it impossible to accurately distinguish the subtle boundaries between normal processing fluctuations and dangerous abnormal signs in real production environments. Since the location and energy proportion of abnormal vibration characteristic frequencies dynamically evolve with changes in the processing material, rotational speed, and feed rate, simple threshold judgment methods are prone to intervention lag, making it impossible for the system to make precise proactive adjustments to process parameters for early anomalies.

[0004] Therefore, existing technologies suffer from low accuracy in identifying anomalies in the early stages of the manufacturing process. Summary of the Invention

[0005] This invention provides an automated control method and system for the production and processing of alloy saw blades, in order to solve the problem of low accuracy in early-stage anomaly identification in the existing technology.

[0006] Firstly, in order to solve the above-mentioned technical problems, the present invention provides an automated control method for the production and processing of alloy saw blades, comprising: Vibration signals during alloy saw blade processing are acquired, and the vibration signals are converted to the frequency domain to determine abnormal frequency distributions. Feature vectors are extracted from the abnormal frequency distribution. If the energy proportion of a specific frequency band in the feature vector exceeds a preset separation threshold, the corresponding feature vector is determined as an abnormal separation vector. Based on the anomaly separation vector, Euclidean distance is calculated and decomposed using a preset support vector machine model, and coupling influencing factors are locked according to a preset fault mechanism topology diagram to determine the coupling influence index. The dominant anomaly source is identified from the coupling influence indicators. If the dominant anomaly source is uneven cutting load, the target speed and target feed rate are adjusted to determine the adjustment scheme. Based on the adjustment scheme, the control command is updated and the feedback vibration signal is detected. If the feedback vibration signal tends to stabilize, an optimization model is constructed and optimization processing is performed to determine the efficiency improvement path. Based on the efficiency improvement path, monitor subsequent signals and perform multi-resolution multi-layer wavelet decomposition to determine multi-scale components. Based on the decay rate of the residual components in the multi-scale components and the downtime cost function, determine the downtime loss reduction threshold. Extract the iteration step size from the downtime loss reduction threshold, determine the target increment, reconstruct the transient stress distribution based on the target increment to determine the coupling variable, perform iterative separation according to the coupling variable, determine the loop condition, and output the minimization control command.

[0007] Preferably, acquiring the vibration signal during alloy saw blade processing, converting the vibration signal to the frequency domain, and determining the abnormal frequency distribution includes: Acquire vibration signals during alloy saw blade processing; The vibration signal is divided into time-domain waveform frames, and the time-domain waveform frames are windowed and subjected to Fourier transform to determine the discrete spectrum. Harmonic components are extracted from the discrete spectrum, and the frequency peaks of the harmonic components are selected to construct a time-frequency feature matrix; Calculate the frequency jump value and amplitude fluctuation rate of each frequency peak in the time-frequency feature matrix. If the frequency jump value or the amplitude fluctuation rate exceeds a preset abnormal threshold, the corresponding feature point is determined as a sudden frequency point. The mutation frequency points are mapped onto the frequency domain spectrum to generate an abnormal frequency distribution.

[0008] Preferably, feature vectors are extracted from the abnormal frequency distribution. If the energy proportion of a specific frequency band in the feature vector exceeds a preset separation threshold, the corresponding feature vector is determined as an abnormal separation vector, including: The abnormal frequency distribution is aggregated to obtain frequency domain energy, and the center frequency and bandwidth span of the frequency domain energy are extracted to construct a feature vector. The energy proportion of a specific frequency band in the feature vector is analyzed. If the energy proportion exceeds a preset separation threshold, the corresponding feature vector is determined as an abnormal separation vector.

[0009] Preferably, based on the anomaly separation vector, Euclidean distance is calculated and decomposed using a preset support vector machine model, and coupling influencing factors are identified according to a preset fault mechanism topology diagram to determine coupling influence indicators, including: The abnormal separation vector is normalized to construct the input matrix; The Euclidean distance between the data points in the input matrix and the preset optimal separating hyperplane is calculated using a preset support vector machine model. The Euclidean distance is then decomposed to identify the deviation dimension components whose deviation exceeds the preset support vector interval. The deviation dimension component is input into a preset fault mechanism topology diagram to lock in the coupling influencing factors; The projection length of the feature vector in the direction of the coupling influencing factor is calculated to obtain the coupling influence index.

[0010] Preferably, the dominant anomaly source is identified from the coupled influence indicators. If the dominant anomaly source is uneven cutting load, the target rotational speed and target feed rate are adjusted to determine the adjustment scheme, including: Principal component extraction is performed on the coupling influence index to determine the dominant anomaly source. If the dominant anomaly source is uneven cutting load, the real-time cutting torque and machine tool spindle stiffness matrix are obtained, and stability calculation is performed to determine the speed limit range. Within the specified speed limit range, the value with the smallest vibration amplitude is selected as the target speed. The target rotational speed is mapped onto a preset stability lobe diagram to determine the critical stable cutting depth corresponding to the target rotational speed; Based on the preset machining efficiency requirements, the target feed rate is determined within the limit range of the critical stable cutting depth; The target rotational speed and the target feed rate are combined to determine the adjustment scheme.

[0011] Preferably, the control command is updated based on the adjustment scheme, and the feedback vibration signal is detected. If the feedback vibration signal tends to stabilize, an optimization model is constructed and optimization processing is performed to determine the efficiency improvement path, including: The adjustment scheme is mapped to G-code instructions to drive the machine tool actuator and collect the feedback vibration signal when the actuator moves. The feedback vibration signal is filtered to obtain a clean signal; Extract characteristic frequencies from the pure signal and calculate the rate of change of the amplitude of the characteristic frequencies; If the amplitude change rate is less than the preset stable threshold, the corresponding adjustment scheme is determined as the steady-state parameter, and the corresponding material removal rate is recorded. An optimization model is constructed based on the steady-state parameters and the material removal rate. The optimization model is then optimized to determine the path to improve efficiency.

[0012] Preferably, based on the efficiency improvement path, subsequent signals are monitored and multi-resolution multi-layer wavelet decomposition is performed to determine multi-scale components. Based on the decay rate of the residual components in the multi-scale components and the downtime cost function, a downtime loss reduction threshold is determined, including: Based on the efficiency improvement path, the acquired signal is discretized and sampled to determine the digital signal; The digital signal is subjected to multi-resolution multi-layer wavelet decomposition to determine multi-scale components, residual components are identified from the multi-scale components, and the decay rate of the residual components is calculated. If the decay rate is greater than the preset convergence threshold, the material utilization rate is calculated to obtain the utilization rate. Calculate the deviation between the utilization rate and the preset benchmark utilization rate to determine the amount of reduction; The reduction is mapped to a preset downtime cost function for quantitative evaluation to determine the threshold for reducing downtime losses.

[0013] Preferably, the iteration step size is extracted from the downtime loss reduction threshold to determine the target increment, the transient stress distribution is reconstructed based on the target increment to determine the coupling variable, iterative separation is performed according to the coupling variable, the loop condition is determined, and a minimization control command is output, including: The original trajectory during the intermittent cutting process is obtained, and the original trajectory is used to extract features to determine the cutting trajectory features; The iteration step size is extracted from the downtime loss reduction threshold to determine the initial increment of the intermittent cutting process. If the peak value of the pulse load in the cutting trajectory features exceeds the preset stress limit, the initial increment is corrected to determine the target increment. Based on the target increment, the transient stress distribution is reconstructed, the compensation vector is calculated, and the compensation vector is subjected to multidimensional mapping processing to determine the coupling variables; Identify the response delay characteristics of the coupled variable; if the response delay characteristics are within a preset stable range, perform iterative separation of high-frequency and low-frequency interference to determine the loop condition. The cyclic condition is injected into the control logic to perform interference hedging within the intermittent cutting coverage area and output a minimize control command.

[0014] Secondly, the present invention provides an automated control system for the production and processing of alloy saw blades, comprising: The signal acquisition module is used to acquire vibration signals during the processing of alloy saw blades, convert the vibration signals to the frequency domain, and determine abnormal frequency distributions. The symptom identification module is used to extract feature vectors from the abnormal frequency distribution. If the energy proportion of a specific frequency band in the feature vector exceeds a preset separation threshold, the corresponding feature vector is determined as an abnormal separation vector. The factor quantification module is used to calculate and decompose Euclidean distance based on the anomaly separation vector using a preset support vector machine model, and to lock the coupling influencing factors and determine the coupling influence index according to the preset fault mechanism topology map. The scheme determination module is used to identify the dominant anomaly source from the coupled influence indicators. If the dominant anomaly source is uneven cutting load, the target speed and target feed rate are adjusted to determine the adjustment scheme. The path optimization module is used to update control commands based on the adjustment scheme and detect feedback vibration signals. If the feedback vibration signals tend to stabilize, an optimization model is constructed and optimization processing is performed to determine the efficiency improvement path. The threshold evaluation module is used to monitor subsequent signals according to the efficiency improvement path and perform multi-resolution multi-layer wavelet decomposition to determine multi-scale components. Based on the decay rate of the residual components in the multi-scale components and the downtime cost function, the downtime loss reduction threshold is determined. The iterative control module is used to extract the iteration step size from the downtime loss reduction threshold, determine the target increment, reconstruct the transient stress distribution based on the target increment to determine the coupling variable, perform iterative separation according to the coupling variable, determine the loop condition, and output the minimization control command.

[0015] Compared with the prior art, the present invention has the following beneficial effects: (1) The present invention performs segmentation, Fourier transform and feature matrix construction on the original vibration signal to obtain the time-varying distribution of abnormal frequency and extract key feature vectors. Through this feature analysis process, the system can separate the key frequency components that characterize the abnormal state from the complex coupled signal generated by high-speed heavy-load cutting, effectively alleviating the problem that the traditional threshold alarm method is difficult to identify early abnormal signs and improving the sensitivity of processing status perception.

[0016] (2) Based on the obtained feature vector, the present invention uses a support vector machine classifier to compare the deviations and combines the fault mechanism topology map to determine the quantitative index of the coupling influencing factors. Through this mapping and analysis from data-driven to physical mechanism, the present invention can transform the abstract signal deviation into specific physical causes such as component fatigue or stiffness reduction, providing numerical basis for the accurate locking of the abnormal source and significantly enhancing the pertinence and timeliness of maintenance.

[0017] (3) The present invention identifies the dominant abnormal source based on the coupling influence index, and determines the cutting parameter adjustment scheme including the target speed and target feed based on the machine tool spindle stiffness matrix and real-time torque data. This scheme reduces the risk of abnormal vibration caused by uneven cutting load by dynamically balancing the fluctuation of cutting force and the machining stability, reduces equipment wear and machining defects caused by parameter mismatch, and ensures the consistency of the machining process.

[0018] (4) The present invention updates the control command by adjusting the cutting parameters and combines wavelet transform to decompose multi-scale components to monitor the changing trend of residual abnormal frequency. This process uses multi-resolution analysis algorithm to capture transient fluctuations in the signal and reduces the threshold by quantifying material utilization and downtime loss, so that managers can intuitively evaluate the effect of production efficiency improvement, effectively reduce unplanned downtime loss and reduce material waste.

[0019] (5) Based on the threshold of downtime loss reduction, the present invention extracts and optimizes the loop conditions, and through iterative application of separation and adjustment process to cover the entire intermittent cutting process, outputs the overall coupling influence factor minimization control command; this strategy realizes dynamic adaptive control of the entire processing process, so that the control system can still maintain efficient and stable interference countermeasure when facing complex cutting trajectory and pulse load, and comprehensively improves the intelligence and economy of alloy saw blade production and processing. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of an automated control method for the production and processing of alloy saw blades provided in the first embodiment of the present invention; Figure 2 This is a schematic diagram of an automated control system for alloy saw blade production and processing provided in the second embodiment of the present invention. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Reference Figure 1 The first embodiment of the present invention provides an automated control method for the production and processing of alloy saw blades, comprising the following steps: S11, acquire the vibration signal during alloy saw blade processing, convert the vibration signal to the frequency domain, and determine the abnormal frequency distribution; S12, extract feature vectors from the abnormal frequency distribution. If the energy proportion of a specific frequency band in the feature vector exceeds a preset separation threshold, then the corresponding feature vector is determined as an abnormal separation vector. S13, Based on the anomaly separation vector, use a preset support vector machine model to perform Euclidean distance calculation and decomposition, and lock the coupling influencing factors according to the preset fault mechanism topology diagram to determine the coupling influence index; S14, Locate the dominant abnormal source from the coupling influence index. If the dominant abnormal source is uneven cutting load, adjust the target speed and target feed rate to determine the adjustment scheme. S15, based on the adjustment scheme, update the control command and detect the feedback vibration signal. If the feedback vibration signal tends to stabilize, construct an optimization model and perform optimization processing to determine the efficiency improvement path. S16, based on the efficiency improvement path, monitor subsequent signals and perform multi-resolution multi-layer wavelet decomposition to determine multi-scale components. Based on the decay rate of the residual components in the multi-scale components and the downtime cost function, determine the downtime loss reduction threshold. S17, extract the iteration step size from the downtime loss reduction threshold, determine the target increment, reconstruct the transient stress distribution based on the target increment to determine the coupling variable, perform iterative separation according to the coupling variable, determine the loop condition, and output the minimization control command.

[0023] In step S11, the vibration signal during the processing of the alloy saw blade is acquired, and the vibration signal is converted to the frequency domain to determine the abnormal frequency distribution, including: Acquire vibration signals during alloy saw blade processing; The vibration signal is divided into time-domain waveform frames, and the time-domain waveform frames are windowed and subjected to Fourier transform to determine the discrete spectrum. Harmonic components are extracted from the discrete spectrum, and the frequency peaks of the harmonic components are selected to construct a time-frequency feature matrix; Calculate the frequency jump value and amplitude fluctuation rate of each frequency peak in the time-frequency feature matrix. If the frequency jump value or the amplitude fluctuation rate exceeds a preset abnormal threshold, the corresponding feature point is determined as a sudden frequency point. The mutation frequency points are mapped onto the frequency domain spectrum to generate an abnormal frequency distribution.

[0024] First, the raw vibration signals during the machining of the carbide saw blade are acquired. In this embodiment, an accelerometer installed near the saw blade body records the physical vibrations generated during the cutting process in real time. The sensor operates at a preset sampling frequency. (e.g., 1000Hz) Acquisition duration is The original vibration sequence (e.g., 10s) forms a sequence containing A time-series data vector of sampling points, where: Secondly, the acquired raw vibration signal is segmented to determine the time-domain waveform frame. To capture short-term non-stationary characteristics during the cutting process, a frame with a length of [missing information] is segmented. The signal sequence is divided into segments according to a fixed step size. Each time-domain waveform frame contains [number] frames. Each data point has a frame length set to 0.5 seconds (i.e., ...). This ensures sufficient frequency resolution when analyzing transient anomalous vibrations and avoids loss of detail in long-sequence analyses.

[0025] Then, windowing and Fourier transform are applied to the time-domain waveform frames to determine the discrete spectrum. To reduce spectral leakage in the frequency domain analysis, the Hanning window function is applied to each frame of the signal. The signal after smoothing and windowing. The calculation formula is as follows: In the formula, For discrete-time indexes within a frame. These are the sampling points of the original waveform after segmentation. This represents the total number of sampling points in a single frame of data. A Fast Fourier Transform is performed on the windowed sequence to transform the signal from the time domain to the frequency domain, obtaining the corresponding discrete spectrum sequence, thereby identifying the energy distribution of each frequency component.

[0026] Subsequently, harmonic components are extracted from the discrete spectrum, and frequency peaks of the harmonic components are selected to construct a time-frequency feature matrix. The system filters out harmonic components characterizing the periodic cutting characteristics of the saw blade (such as peaks corresponding to 50Hz, 100Hz, and 150Hz) from the spectrum of each frame, and records the frequency values ​​and amplitudes of each harmonic at different time frames. The main harmonic features extracted from the time frames are arranged to construct a time-frequency feature matrix to characterize the dynamic trend of frequency evolution over time.

[0027] Furthermore, the frequency jump values ​​and amplitude fluctuation rates of each frequency peak in the time-frequency characteristic matrix are calculated. Frequency jump values The calculation logic is as follows: Amplitude volatility The calculation logic is as follows: In the formula, and These represent the peak frequency and amplitude of the current frame, respectively. and These represent the corresponding parameters from the previous frame. If the frequency jump value... Or amplitude volatility If the frequency jump exceeds a preset abnormal threshold, the corresponding feature point is identified as a sudden change frequency point. In this embodiment, the preset frequency jump threshold is set to 10Hz, and the amplitude fluctuation threshold is set to 20%. These preset values ​​are determined based on multiple sets of experimental conclusions under high-speed heavy-load cutting conditions (such as a rotational speed in the range of 1000r / min to 1200r / min), and can effectively distinguish between normal fluctuations caused by uneven workpiece material and sudden abnormal changes caused by saw blade damage.

[0028] Finally, the identified abrupt change frequency points are mapped onto the frequency domain spectrum for localization, generating an anomalous frequency distribution. By visually marking the coordinates (time and frequency) of the abrupt change points on the frequency domain spectrum, energy abrupt changes, such as those occurring at 120Hz at the 5th second, can be clearly identified. This anomalous frequency distribution intuitively demonstrates the specific time and frequency span of the anomalous occurrence, providing accurate data support for subsequent inferences about saw blade wear or process parameter mismatches.

[0029] In step S12, feature vectors are extracted from the abnormal frequency distribution. If the energy proportion of a specific frequency band in the feature vector exceeds a preset separation threshold, the corresponding feature vector is determined as an abnormal separation vector, including: The abnormal frequency distribution is aggregated to obtain frequency domain energy, and the center frequency and bandwidth span of the frequency domain energy are extracted to construct a feature vector. The energy proportion of a specific frequency band in the feature vector is analyzed. If the energy proportion exceeds a preset separation threshold, the corresponding feature vector is determined as an abnormal separation vector.

[0030] First, the anomalous frequency distribution is aggregated to determine the frequency domain energy. The anomalous frequency distribution generated in step S11 is integrated in the frequency domain to obtain the frequency domain energy distribution reflecting the change of signal energy with frequency. The center frequency is then extracted from this distribution. With bandwidth span Center frequency The calculation formula is as follows: In the formula, The frequency value of the abnormal frequency point. This represents the energy intensity corresponding to that frequency point. This represents the total number of anomalous frequency points. The center frequency, measured in Hz, characterizes the central location of the anomalous signal energy distribution. Bandwidth span. The frequency range covered by the anomalous energy distribution is defined by the following formula: In the formula, This represents the upper frequency limit of the anomalous energy distribution. The lower limit frequency is determined by the extracted center frequency. With bandwidth span Combine them to construct feature vectors This feature vector is used to digitally characterize the vibration characteristics of the alloy saw blade during the cutting process.

[0031] Secondly, the energy proportion of a specific frequency band in the feature vector is analyzed. In this embodiment, the specific frequency band refers to a pre-determined frequency range (e.g., 200Hz to 400Hz) that is highly correlated with the wear or chipping characteristics of the alloy saw blade teeth. The energy proportion is determined by accumulating the frequency domain energy within this range and calculating its percentage in the total frequency domain energy. Energy percentage The calculation formula is as follows: In the formula, and These are the start and end frequencies of a specific frequency band. Let be the frequency domain energy density function. and These are the frequency boundaries of the entire frequency domain.

[0032] Then, the energy percentage The value is compared with a preset separation threshold. In this embodiment, the preset separation threshold is set to 35%. This threshold is set based on fatigue wear test data during the intermittent cutting process of the alloy saw blade. Experiments show that when the energy weight in the 200Hz to 400Hz frequency band exceeds 35%, the saw blade vibration mode will change significantly, which usually means that the saw teeth have entered an early fatigue damage state.

[0033] Finally, if the energy percentage If the separation threshold is exceeded, the corresponding feature vector is identified as an anomaly separation vector. This feature vector is then assigned an anomaly attribute identifier, serving as the triggering logic for subsequent support vector machine classification and coupling effect analysis. This step enables accurate separation of early abnormal states from massive vibration data, providing a basis for proactive intervention in processing stability.

[0034] In step S13, based on the anomaly separation vector, Euclidean distance is calculated and decomposed using a preset support vector machine model. Coupled influencing factors are then identified according to a preset fault mechanism topology diagram, and coupling influence indices are determined, including: The abnormal separation vector is normalized to construct the input matrix; The Euclidean distance between the data points in the input matrix and the preset optimal separating hyperplane is calculated using a preset support vector machine model. The Euclidean distance is then decomposed to identify the deviation dimension components whose deviation exceeds the preset support vector interval. The deviation dimension component is input into a preset fault mechanism topology diagram to lock in the coupling influencing factors; The projection length of the feature vector in the direction of the coupling influencing factor is calculated to obtain the coupling influence index.

[0035] First, a normal mode training set was obtained, and a support vector machine (SVM) was trained on this set to establish an SVM classification model. The training data consisted of over 500 sets of stable cutting signal features from alloy saw blades acquired at constant rotational speed (1100 r / min) and feed rate (0.05 mm / tooth). During training, a Gaussian radial basis function was used as the kernel function to map the input data to a high-dimensional feature space, and the weight vector was determined by solving a dual optimization problem. and bias terms Thus, the optimal separating hyperplane is determined: Simultaneously, the geometric margin from the support vectors to the optimal separating hyperplane is calculated to determine the support vector margin. This interval represents the maximum permissible range of random disturbances under normal cutting conditions, and is set to 0.15 in this embodiment.

[0036] Secondly, based on the anomaly separation vector, the corresponding feature vectors are normalized to determine the input matrix. To eliminate the interference of different physical quantity dimensions on the calculation accuracy, the standard deviation normalization method is used to normalize the extracted center frequency. With bandwidth span Processing is performed. The standardized components are then processed. The calculation formula is as follows: In the formula, These are the original component values. This is the arithmetic mean of the corresponding dimension in the normal mode training set. represents the standard deviation of the corresponding dimension. The processed feature vectors form the input matrix, and their values ​​are distributed within the standard normal distribution interval.

[0037] Then, the Euclidean distance between the data points in the input matrix and the optimal separating hyperplane is calculated using a support vector machine classification model. (Euclidean distance) The calculation formula is as follows: In the formula, This is the normalized feature vector corresponding to the current abnormal symptoms. This represents the norm of the weight vector. Subsequently, the Euclidean distance is decomposed to identify deviations exceeding the support vector margin. The deviation dimension component. If the component deviation in the frequency dimension... satisfy If so, it is determined to be a deviation dimension component caused by the change in cutting state.

[0038] Subsequently, the off-dimensional components are input into a pre-defined fault mechanism topology diagram to lock in the coupling influencing factors. The fault mechanism topology diagram is essentially a multi-dimensional mapping matrix. This method is used to quantify the causal relationship between signal characteristic perturbations and physical failure mechanisms. The construction process involves collecting historical failure case data during the machining of alloy saw blades, covering four typical working conditions: saw tooth wear, matrix fatigue, uneven cutting load, and lubrication failure. The method then extracts the signal characteristic deviation values ​​of each sample relative to the normal state, including the center frequency offset. Bandwidth variation and energy percentage fluctuation The correlation analysis method is used to calculate the sensitivity influence weights between each characteristic deviation value and a specific physical cause, and a mapping matrix is ​​constructed. : In the formula, Representing the The first physical reason for the first The sensitivity of signal feature deviations affects the weights. In this embodiment, These correspond to four typical failure mechanisms: sawtooth wear, matrix fatigue, uneven cutting load, and lubrication failure. These correspond to the center frequency, bandwidth, and energy percentage characteristics, respectively. Taking the first row of the matrix as an example, This indicates the weight of the effect of sawtooth wear on the center frequency deviation. This indicates the weight of the impact of sawtooth wear on bandwidth change. This represents the weight of the impact of sawtooth wear on the energy percentage fluctuation. Weight value The value ranges from 0 to 1, with larger values ​​indicating greater sensitivity to the physical cause. For example, experimental calibration results show the weight of uneven cutting load on center frequency shift. Set to 0.85. Calculate the overall contribution score for each physical cause: The item with the highest score is selected as the dominant inducing factor. This includes the deviation dimension component obtained from the center frequency decomposition in the observation vector. When in a dominant position, through the matrix The weighted calculation yields the contribution score of uneven cutting load. The value is the highest, thus identifying the current coupling influencing factor as uneven cutting load.

[0039] Finally, the projection length of the eigenvector onto the direction of the coupling influencing factors is calculated to determine the coupling influence index. The normalized eigenvector is then... The unit direction vector of the coupling influencing factors in the feature space Perform a dot product operation to calculate the projected length:

[0040] This indicator As a dimensionless scalar, it reflects the severity of the interference caused by specific physical factors on the current anomalous signal. If... The continuous increase in the indicator indicates an intensified coupling effect, necessitating subsequent parameter adjustment procedures. This step achieves a quantitative characterization of the cause of the anomaly, providing a parameter basis for precise control.

[0041] In step S14, the dominant anomaly source is identified from the coupling influence indicators. If the dominant anomaly source is uneven cutting load, the target rotational speed and target feed rate are adjusted to determine the adjustment scheme, including: Principal component extraction is performed on the coupling influence index to determine the dominant anomaly source. If the dominant anomaly source is uneven cutting load, the real-time cutting torque and machine tool spindle stiffness matrix are obtained, and stability calculation is performed to determine the speed limit range. Within the specified speed limit range, the value with the smallest vibration amplitude is selected as the target speed. The target rotational speed is mapped onto a preset stability lobe diagram to determine the critical stable cutting depth corresponding to the target rotational speed; Based on the preset machining efficiency requirements, the target feed rate is determined within the limit range of the critical stable cutting depth; The target rotational speed and the target feed rate are combined to determine the adjustment scheme.

[0042] First, principal component analysis (PCA) is performed on the coupling impact indicators to identify the dominant anomaly source. A sliding time window containing 20 consecutive sampling points is extracted to obtain the coupling impact indicators at each sampling point, encompassing three dimensions: center frequency, bandwidth span, and energy percentage. A system of [size missing] is then constructed. Observation feature set matrix .matrix Each column vector in represent The instantaneous coupling state at time t. For the matrix Perform a centering process, calculate the mean vector for each dimension, and subtract the mean from the original observations to obtain the centered matrix. Then, the covariance matrix of the set is calculated, and the covariance matrix is... Perform eigenvalue decomposition by solving the characteristic equation: Extracting feature values and the corresponding unit eigenvector Sort the eigenvalues ​​in descending order of their numerical values ​​and calculate the cumulative contribution rate: Select the one that satisfies The smallest integer The corresponding feature vector is used as the principal component direction. In this embodiment, if the contribution rate of the first principal component reaches 85% or more, then the feature vector is extracted. As the dominant vector characterizing the evolution trend of the current processing state, the extracted principal component direction vector is used. Compared with the preset physical cause vector in the fault mechanism topology diagram Perform cosine similarity calculation. These are standard feature vectors in the fault mechanism topology diagram that correspond to causes such as uneven cutting load, sawtooth wear, or matrix fatigue. Through traversal comparison, if the highest similarity value points to the uneven cutting load vector and the score exceeds the preset matching threshold of 0.8, then the dominant anomaly source is identified as uneven cutting load. This process utilizes statistical features to suppress single-time random interference, ensuring the accuracy of anomaly source identification.

[0043] Secondly, if the dominant anomaly source is confirmed to be uneven cutting load, then the real-time cutting torque and machine tool spindle stiffness matrix are obtained, and stability calculations are performed to determine the speed limit range. Machine tool spindle stiffness matrix It was determined in advance through static loading tests on the machine tool spindle. Matrix, unit This is used to characterize the structural rigidity of the spindle in the x, y, and z directions. The real-time cutting torque is obtained by combining the output current of the machine tool spindle power sensor with a power-torque conversion model, and the unit is... Dynamic stability is determined using analytical methods: In the formula, This indicates the critical stable cutting depth. This indicates the cutting force coefficient related to the alloy saw blade material. Indicates the total number of teeth on the saw blade. This represents the real part of the frequency response function of the spindle system. By searching within the preset spindle speed range, a function that satisfies this requirement is... For frequencies greater than the current cutting depth, identify the set of frequencies that will not cause chatter, i.e., the speed limit range. The unit is For conventional carbide saw blade processing, this range typically extends between 900 r / min and 1400 r / min.

[0044] Then, within the speed limit range, the value with the smallest vibration amplitude is selected as the target speed. The system performs discrete sampling within the speed limit range in steps of 20 r / min, and combines this with the discrete spectrum data from step S11 to predict the dynamic response of the spindle system at each speed point. The speed point with the smallest predicted vibration displacement amplitude is selected as the target speed. The selection is based on utilizing the structural dynamics characteristics to avoid the resonant region of the system's natural frequency, thereby suppressing forced vibrations excited by load fluctuations at the physical source.

[0045] Subsequently, the target rotational speed is mapped onto a pre-defined stability lobe diagram to determine the critical stable cutting depth corresponding to the target rotational speed. The stability lobe diagram is a stability region distribution map pre-established by obtaining the frequency response function through a hammer test on the tip of the carbide saw blade and combining it with the cutting force dynamics equation. The target rotational speed is then retrieved from the diagram. The highest point of the corresponding leaf blade curve determines the maximum depth of cut that can maintain stability at the current rotational speed, i.e., the critical stable depth of cut. The unit is mm.

[0046] Furthermore, based on the preset machining efficiency requirements, the target feed rate is determined within the limit of the critical stable depth of cut. Machining efficiency requirements. This refers to the required material removal rate per unit time. Target feed rate. The method for determining it is as follows: In the formula, Represents the feed per tooth, in units of The system performs a limiting process on the calculated feed rate to ensure that it is not less than 0.02 mm / tooth (to ensure chip formation) and not more than 0.15 mm / tooth (to prevent the serration structure from breaking).

[0047] Finally, the target speed and target feed rate are combined to determine the adjustment scheme. This scheme forms a set of control parameters including speed commands and feed rate correction coefficients, which are sent to the execution buffer of the digital control system to drive the frequency converter and servo drive to adjust the output synchronously. This process, through quantitative analysis of the physical limitations of machining, achieves targeted compensation for uneven load anomalies, ensuring the smoothness of the saw blade machining process.

[0048] In step S15, the control command is updated based on the adjustment scheme, and the feedback vibration signal is detected. If the feedback vibration signal tends to stabilize, an optimization model is constructed and optimization processing is performed to determine the efficiency improvement path, including: The adjustment scheme is mapped to G-code instructions to drive the machine tool actuator and collect the feedback vibration signal when the actuator moves. The feedback vibration signal is filtered to obtain a clean signal; Extract characteristic frequencies from the pure signal and calculate the rate of change of the amplitude of the characteristic frequencies; If the amplitude change rate is less than the preset stable threshold, the corresponding adjustment scheme is determined as the steady-state parameter, and the corresponding material removal rate is recorded. An optimization model is constructed based on the steady-state parameters and the material removal rate. The optimization model is then optimized to determine the path to improve efficiency.

[0049] First, the adjustment scheme is mapped to G-code instructions to drive the machine tool actuator. The system converts the target speed and target feed rate determined in step S14 into a text instruction stream recognizable by the CNC system, for example, by generating G01 machining code containing the spindle speed instruction "S" and the feed rate instruction "F" through string formatting operations. After receiving the code instructions, the actuator synchronously adjusts the output frequency of the spindle frequency converter and the feed axis servo motor, so that the carbide saw blade enters the cutting state according to the new process parameters. During this process, feedback vibration signals during the actuator's operation are collected by a three-dimensional accelerometer mounted on the spindle bearing housing.

[0050] Secondly, the feedback vibration signal is filtered to obtain a clean signal. To eliminate low-frequency environmental noise generated by auxiliary equipment such as the machine tool coolant pump and chip conveyor, a fourth-order Butterworth bandpass filter is used to process the feedback vibration signal. The lower cutoff frequency of the filter is set to 50Hz, and the upper cutoff frequency is set to 1000Hz. This frequency band covers all effective cutting vibration characteristics of the carbide saw blade from the fundamental frequency to higher harmonics. The processed signal effectively suppresses out-of-band interference, forming a clean signal that reflects the actual changes in cutting load.

[0051] Then, characteristic frequencies are extracted from the pure signal, and the rate of change of the amplitude of the characteristic frequencies is calculated. By performing a short-time Fourier transform on the pure signal, the frequency component with the highest energy peak is identified as the characteristic frequency. Calculate the fluctuation of the amplitude corresponding to this characteristic frequency over three consecutive observation periods. Amplitude change rate. The calculation formula is as follows: In the formula, For the first The amplitude of the characteristic frequency within each sampling period, for The average amplitude of each sampling period. The value is 10. This index is used to quantitatively evaluate the convergence of the cutting process after parameter adjustment.

[0052] Subsequently, if the rate of change of amplitude If the material removal rate is less than the preset stability threshold, the corresponding adjustment scheme is determined as the steady-state parameter, and the corresponding material removal rate is recorded. In this embodiment, the preset stability threshold is set to 5%. This value is determined based on the wear resistance test of the saw blade coating. When the amplitude fluctuation is less than 5%, the cutting heat and mechanical stress reach a state of equilibrium. At this time, the current rotational speed and feed rate are locked as steady-state parameters, and the material removal rate is calculated. Material removal rate The calculation formula is as follows: In the formula, For steady-state speed, This refers to the number of saw blade teeth. The feed per tooth. For cutting depth, The saw blade cut width, in units of .

[0053] Finally, an optimization model is constructed based on steady-state parameters and material removal rate. This model is then optimized to determine the path to improve efficiency. An optimization model is constructed with the objective function of maximizing material removal rate and constraints of machine tool spindle power and surface roughness. The objective function is expressed as: in and The pre-set weighting system is based on balancing output per unit time and equipment protection life. Through cutting tests on workpieces of different materials, the weights of material removal rate and vibration amplitude on the total processing cost are obtained. When pursuing high output, the weights are increased... Value (e.g., set to 0.75); when processing high-value materials, vibration must be strictly controlled to ensure surface quality, increase... The value is set (e.g., to 0.8) to ensure the convergence of the optimization process under specific production targets. The specific operation of the optimization process is as follows: first, the steady-state parameter is used as the initial iteration point; then, the rotational speed is finely adjusted along the gradient ascent direction. With feed rate Next, it is determined whether the adjusted parameters still meet the aforementioned stability threshold requirements. If they do, the iterative action continues until the objective function reaches a local optimum. Finally, the parameter evolution trajectory obtained after the iteration is determined as the efficiency improvement path. This path indicates the parameter adjustment sequence that transitions from the current steady state to a high-efficiency machining state while maintaining cutting stability.

[0054] In step S16, following the efficiency improvement path, subsequent signals are monitored and multi-resolution multi-layer wavelet decomposition is performed to determine multi-scale components. Based on the attenuation rate of the residual components in the multi-scale components and the downtime cost function, a downtime loss reduction threshold is determined, including: Based on the efficiency improvement path, the acquired signal is discretized and sampled to determine the digital signal; The digital signal is subjected to multi-resolution multi-layer wavelet decomposition to determine multi-scale components, residual components are identified from the multi-scale components, and the decay rate of the residual components is calculated. If the decay rate is greater than the preset convergence threshold, the material utilization rate is calculated to obtain the utilization rate. Calculate the deviation between the utilization rate and the preset benchmark utilization rate to determine the amount of reduction; The reduction is mapped to a preset downtime cost function for quantitative evaluation to determine the threshold for reducing downtime losses.

[0055] Since the efficiency improvement path determined in step S15 is a predictive model based on prior steady-state data, directly applying this path in its entirety may encounter nonlinear disturbances caused by uneven material structure or intermittent cutting impacts during actual processing. Therefore, step S16, as a technical verification and economic evaluation step, aims to confirm whether the process optimization has truly achieved energy convergence by performing real-time stability checks on subsequent signals after parameter adjustments. This establishes an evaluation index from physical signals to economic gains and losses, providing closed-loop feedback for subsequent iterative iterations.

[0056] First, signals are acquired based on the efficiency improvement path, and subsequent signals are discretized to determine the digital signals. The actuator enters the machining section according to the optimized rotational speed and feed rate. During the cutting process, a high-sensitivity vibration sensor captures the vibration waveform of the carbide saw blade. The sampling frequency is set to 5000Hz, and the sampling bit depth is set to 16 bits. The continuous signal is converted into a discrete digital signal sequence through an analog-to-digital conversion process. This sampling frequency is selected based on the distribution pattern of intermittent cutting high-frequency impact characteristics within 2000Hz, aiming to fully cover the dynamic response of the saw blade at the moment of entry and exit, and avoid sampling aliasing.

[0057] Secondly, multi-resolution, multi-level wavelet decomposition is performed on the digital signal to determine multi-scale components. The Daubechies4 wavelet basis function is used to perform four-level discrete wavelet decomposition on the digital signal sequence. The decomposition process breaks down the signal into approximate components reflecting the rotational fundamental frequency characteristics and detail components reflecting transient impact characteristics. Through this time-frequency analysis method, the original signal is transformed into a set of multi-scale components capable of characterizing processing stability in multiple dimensions.

[0058] Then, residual components are identified from the multi-scale components, and their decay rates are calculated. The detail components of the first and second layers in the decomposition results are extracted as residual components; these components contain residual system oscillations caused by parameter variations. The specific process for calculating the decay rate involves obtaining the residual component amplitude envelope values ​​at two different observation times, calculating the natural logarithm of the ratio of the two amplitude envelope values, and then dividing this logarithm by the time difference between the two observation times to obtain the decay rate in units of 1 / ms. This rate quantitatively measures the speed at which anomalous energy dissipates over time.

[0059] Subsequently, if the decay rate exceeds a preset convergence threshold, material utilization is calculated to obtain the utilization rate. The preset convergence threshold is set to 0.15ms. -1 This value is determined based on the damping characteristics of the alloy saw blade matrix material at standard cutting temperatures, representing the minimum criterion for the system to transition from fluctuation to steady state. Specifically, based on the dynamic damping test experiment of the alloy saw blade matrix material at standard cutting temperatures (20℃±5℃), the system attenuation curve is measured by hammer impact method, and the statistical lower limit of the stable attenuation rate range is taken to ensure coverage of more than 95% of working conditions. Regarding the rationality of the utilization calculation logic, material utilization is obtained by comparing effective processing output with actual resource consumption. First, the effective material removal volume is calculated, which is the theoretical cutting volume that meets the workpiece design dimensions and surface accuracy requirements. Then, the actual total processing volume is calculated by monitoring the actual movement trajectory of the saw blade into the workpiece using a laser displacement sensor and combining it with feed axis pulse data. The specific process for calculating utilization is: calculating the ratio of effective material removal volume to actual total processing volume. In this definition, the actual total processing volume includes not only normal chips but also chipping caused by vibration instability, overcutting losses due to trajectory deviation, and the volume of scrap generated. This calculation method allows the utilization rate to truly reflect the reduction in material waste due to improved cutting stability, solving the problem that simply relying on wear amount cannot reflect the loss of machining quality.

[0060] Furthermore, the deviation between the utilization rate and the preset benchmark utilization rate is calculated to determine the reduction amount. The benchmark utilization rate is set at 85%, which is derived from the historical average qualified output statistics of the same model of saw blade when processed using traditional fixed process parameters. The reduction amount, from a technical perspective, characterizes the degree of reduction in ineffective resource consumption achieved after the application of efficiency improvement paths due to vibration suppression and improved cutting trajectory accuracy.

[0061] Finally, the reduction is mapped to a preset downtime cost function for quantitative evaluation to determine the downtime loss reduction threshold. The downtime cost function is a nonlinear mapping model used to evaluate the stability and continuity of the processing. The model construction process includes statistically analyzing the energy load intensity per unit time of the alloy saw blade production line, equipment maintenance cycle fluctuations, and tool life attenuation rates; analyzing the technical correlation between abnormal vibration intensity and unplanned downtime probability in historical production data; and establishing a mapping relationship between vibration energy dissipation rate and processing continuity. The calculated reduction is used as an input variable, input into the downtime cost function, processed by the numerical mapping model within the function, and the corresponding value is output to determine the downtime loss reduction threshold.

[0062] In step S17, the iteration step size is extracted from the downtime loss reduction threshold to determine the target increment. Based on the target increment, the transient stress distribution is reconstructed to determine the coupling variables. According to the coupling variables, iterative separation is performed to determine the loop conditions and output the minimization control command, including: The original trajectory during the intermittent cutting process is obtained, and the original trajectory is used to extract features to determine the cutting trajectory features; The iteration step size is extracted from the downtime loss reduction threshold to determine the initial increment of the intermittent cutting process. If the peak value of the pulse load in the cutting trajectory features exceeds the preset stress limit, the initial increment is corrected to determine the target increment. Based on the target increment, the transient stress distribution is reconstructed, the compensation vector is calculated, and the compensation vector is subjected to multidimensional mapping processing to determine the coupling variables; Identify the response delay characteristics of the coupled variable; if the response delay characteristics are within a preset stable range, perform iterative separation of high-frequency and low-frequency interference to determine the loop condition. The cyclic condition is injected into the control logic to perform interference hedging within the intermittent cutting coverage area and output a minimize control command.

[0063] First, the original trajectory during the intermittent cutting process is acquired, and features are extracted from the original trajectory to determine the cutting trajectory characteristics. The spatial coordinate changes of the carbide saw blade at the instants of entry and exit from the workpiece are captured using the data stream fed back from the machine tool's linear encoder. A polynomial fitting method is used to smooth the acquired coordinate data, extracting the instantaneous displacement deviation and velocity fluctuation amplitude of the tool tip, and constructing a cutting trajectory feature set reflecting the spatial geometric properties.

[0064] Secondly, the iteration step size is extracted from the downtime loss reduction threshold to determine the initial increment of the intermittent cutting process. The downtime loss reduction threshold, as a dynamic technical indicator output in step S16, characterizes the technical contribution of the current optimized path to maintaining the continuity of system operation. Through a preset step size mapping function, the downtime loss reduction threshold is converted into the adjustment range of the control parameters.

[0065] The construction process of the step size mapping function involves performing intermittent cutting calibration experiments under various feed gradients, recording the correlation data between the parameter variation step size and the vibration energy decay rate during the recovery of the machining system from a disturbed state to a steady state; and using linear regression to fit the correlation data to determine the proportional mapping relationship between the downtime loss reduction threshold and the iteration step size. This function is set based on the dynamic response bandwidth of the machine tool actuator and the structural stiffness of the alloy saw blade substrate, used to limit the step size adjustment to not exceed the physical tracking limit of the servo drive system. When the downtime loss reduction threshold is in a high value range greater than 0.7, a larger iteration step size is allocated to accelerate the parameter search speed; when the downtime loss reduction threshold is in a low value range less than 0.3, a smaller iteration step size is allocated. The initial increment is determined as the scalar representation of this iteration step size in the current feed direction.

[0066] Then, if the peak value of the pulse load in the cutting trajectory characteristics exceeds the preset stress limit, the initial increment is corrected to determine the target increment. The preset stress limit is set to 450 MPa, which is derived based on the yield strength of the alloy saw blade matrix material and the fatigue safety factor during intermittent cutting, representing the critical load point at which the matrix does not undergo permanent plastic deformation. When the peak value of the pulse load is detected to exceed this limit, a forced correction is performed by reducing the amplitude of the initial increment to obtain the target increment that meets the physical safety constraints.

[0067] Furthermore, based on the target increment, the transient stress distribution is reconstructed, the compensation vector is calculated, and the compensation vector is subjected to multidimensional mapping processing to determine the coupling variables. A pre-set equivalent mechanical model is driven by the target increment to simulate the dynamic stress field of the saw blade tooth tip in the cutting zone. By extracting the gradient change trend of each node in the stress field, the compensation vector used to offset trajectory deviations is calculated. The multidimensional mapping process involves projecting the compensation vector onto a multidimensional control space composed of spindle speed, feed per tooth, and depth of cut. The coupling variables that can synchronously adjust multiple actuators are calculated through the correlation matrix, thereby establishing a technical connection between micro-vibration compensation and macro-process parameters.

[0068] Subsequently, the response delay characteristics of the coupled variables are identified. If the response delay characteristics are within a preset stable range, iterative separation is performed on high-frequency and low-frequency interference to determine the loop conditions. The preset stable range is set to 5ms to 15ms. The lower limit of 5ms is used to eliminate inherent communication jitter in the system, and its setting is based on 5 times the 1ms scan cycle of the machine tool controller; the upper limit of 15ms is used to maintain phase synchronization between control compensation and the physical vibration field, and its setting is based on ensuring that the lag angle of the compensation signal does not exceed 90 degrees. The response delay characteristics are determined by calculating the difference between the timestamp of the controller issuing the command and the timestamp of the grating ruler's feedback displacement.

[0069] If the time delay is within the preset stable range, iterative separation is performed. A low-pass filter is used to filter out low-frequency servo interference caused by sudden load changes, and a high-frequency notch filter is used to identify and extract high-frequency impact interference generated by self-excited vibration of the cutting process. By comparing the residual energy of the interference components, the cyclic conditions that satisfy the system convergence requirements are determined.

[0070] Finally, the loop condition is injected into the control logic to perform disturbance cancellation within the full coverage of intermittent cutting, outputting a minimized control command. The disturbance characteristics are converted into servo axis position command offsets, and the original feed command is dynamically corrected within the real-time interpolation cycle of the CNC system. This operation achieves dynamic cancellation of transient fluctuations by inverting the characteristics of the disturbance components and superimposing them into the control loop, ultimately generating a minimized control command sent to the servo driver, achieving closed-loop dynamic optimization of cutting parameters while maintaining machining continuity.

[0071] In summary, this invention discloses an automated control method for alloy saw blade production and processing. It acquires the original vibration signal and converts it into a frequency domain spectrum to determine the abnormal frequency distribution; extracts feature vectors and determines the abnormal separation vector when the energy percentage exceeds the limit; uses a support vector machine to calculate the deviation between the feature vector and the normal mode, and combines this with a fault mechanism topology diagram to determine the coupling influence index; adjusts the rotational speed and feed rate based on the dominant abnormal source and machine tool stiffness characteristics to determine the adjustment scheme and update the equipment control commands; and uses wavelet transform to analyze the multi-scale components of the signal, and achieves minimized control of the coupling influence throughout the intermittent cutting process by iteratively optimizing the cyclic conditions, thus addressing the problem of low accuracy in early-stage abnormality identification in existing technologies.

[0072] Reference Figure 2 The second embodiment of the present invention provides an automated control system for the production and processing of alloy saw blades, comprising: The signal acquisition module is used to acquire vibration signals during the processing of alloy saw blades, convert the vibration signals to the frequency domain, and determine abnormal frequency distributions. The symptom identification module is used to extract feature vectors from the abnormal frequency distribution. If the energy proportion of a specific frequency band in the feature vector exceeds a preset separation threshold, the corresponding feature vector is determined as an abnormal separation vector. The factor quantification module is used to calculate and decompose Euclidean distance based on the anomaly separation vector using a preset support vector machine model, and to lock the coupling influencing factors and determine the coupling influence index according to the preset fault mechanism topology map. The scheme determination module is used to identify the dominant anomaly source from the coupled influence indicators. If the dominant anomaly source is uneven cutting load, the target speed and target feed rate are adjusted to determine the adjustment scheme. The path optimization module is used to update control commands based on the adjustment scheme and detect feedback vibration signals. If the feedback vibration signals tend to stabilize, an optimization model is constructed and optimization processing is performed to determine the efficiency improvement path. The threshold evaluation module is used to monitor subsequent signals according to the efficiency improvement path and perform multi-resolution multi-layer wavelet decomposition to determine multi-scale components. Based on the decay rate of the residual components in the multi-scale components and the downtime cost function, the downtime loss reduction threshold is determined. The iterative control module is used to extract the iteration step size from the downtime loss reduction threshold, determine the target increment, reconstruct the transient stress distribution based on the target increment to determine the coupling variable, perform iterative separation according to the coupling variable, determine the loop condition, and output the minimization control command.

[0073] It should be noted that the automated control system for alloy saw blade production and processing provided in this embodiment of the invention is used to execute all process steps of the automated control method for alloy saw blade production and processing described above. The working principles and beneficial effects of the two are one-to-one, so they will not be described again.

[0074] This invention also provides an electronic device. The electronic device includes a processor, a memory, and a computer program stored in the memory and executable on the processor, such as an automated control program for the production and processing of alloy saw blades. When the processor executes the computer program, it implements the steps described in the various embodiments of the automated control method for the production and processing of alloy saw blades, for example... Figure 1 The step S11 shown. Alternatively, when the processor executes the computer program, it implements the functions of each module in the above system embodiments, such as the data acquisition module.

[0075] For example, the computer program may be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the electronic device.

[0076] The electronic device may be a desktop computer, laptop, handheld computer, or smart tablet, etc. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above components are merely examples of electronic devices and do not constitute a limitation on the electronic device. It may include more or fewer components than described above, or combine certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, etc.

[0077] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the electronic device, connecting all parts of the electronic device via various interfaces and lines.

[0078] The memory can be used to store the computer program or module. The processor implements various functions of the electronic device by running or executing the computer program or module stored in the memory and calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0079] If the modules integrated into the electronic device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or system capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electrical carrier signals and telecommunication signals.

[0080] It should be noted that the system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the system embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0081] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. An automated control method for the production and processing of alloy saw blades, characterized in that, include: Vibration signals during alloy saw blade processing are acquired, and the vibration signals are converted to the frequency domain to determine abnormal frequency distributions. Feature vectors are extracted from the abnormal frequency distribution. If the energy proportion of a specific frequency band in the feature vector exceeds a preset separation threshold, the corresponding feature vector is determined as an abnormal separation vector. Based on the anomaly separation vector, Euclidean distance is calculated and decomposed using a preset support vector machine model, and coupling influencing factors are locked according to a preset fault mechanism topology diagram to determine the coupling influence index. The dominant anomaly source is identified from the coupling influence indicators. If the dominant anomaly source is uneven cutting load, the target speed and target feed rate are adjusted to determine the adjustment scheme. Based on the adjustment scheme, the control command is updated and the feedback vibration signal is detected. If the feedback vibration signal tends to stabilize, an optimization model is constructed and optimization processing is performed to determine the efficiency improvement path. Based on the efficiency improvement path, monitor subsequent signals and perform multi-resolution multi-layer wavelet decomposition to determine multi-scale components. Based on the decay rate of the residual components in the multi-scale components and the downtime cost function, determine the downtime loss reduction threshold. Extract the iteration step size from the downtime loss reduction threshold, determine the target increment, reconstruct the transient stress distribution based on the target increment to determine the coupling variable, perform iterative separation according to the coupling variable, determine the loop condition, and output the minimization control command.

2. The automated control method for alloy saw blade production and processing according to claim 1, characterized in that, The process of acquiring vibration signals during alloy saw blade processing, converting the vibration signals to the frequency domain, and determining abnormal frequency distributions includes: Acquire vibration signals during alloy saw blade processing; The vibration signal is divided into time-domain waveform frames, and the time-domain waveform frames are windowed and subjected to Fourier transform to determine the discrete spectrum. Harmonic components are extracted from the discrete spectrum, and the frequency peaks of the harmonic components are selected to construct a time-frequency feature matrix; Calculate the frequency jump value and amplitude fluctuation rate of each frequency peak in the time-frequency feature matrix. If the frequency jump value or the amplitude fluctuation rate exceeds a preset abnormal threshold, the corresponding feature point is determined as a sudden frequency point. The mutation frequency points are mapped onto the frequency domain spectrum to generate an abnormal frequency distribution.

3. The automated control method for alloy saw blade production and processing according to claim 1, characterized in that, The step of extracting feature vectors from the abnormal frequency distribution, wherein if the energy proportion of a specific frequency band in the feature vector exceeds a preset separation threshold, then the corresponding feature vector is determined as an abnormal separation vector, includes: The abnormal frequency distribution is aggregated to obtain frequency domain energy, and the center frequency and bandwidth span of the frequency domain energy are extracted to construct a feature vector. The energy proportion of a specific frequency band in the feature vector is analyzed. If the energy proportion exceeds a preset separation threshold, the corresponding feature vector is determined as an abnormal separation vector.

4. The automated control method for alloy saw blade production and processing according to claim 1, characterized in that, Based on the anomaly separation vector, a preset support vector machine model is used to calculate and decompose Euclidean distance, and coupling influencing factors are identified according to a preset fault mechanism topology diagram to determine coupling influence indicators, including: The abnormal separation vector is normalized to construct the input matrix; The Euclidean distance between the data points in the input matrix and the preset optimal separating hyperplane is calculated using a preset support vector machine model. The Euclidean distance is then decomposed to identify the deviation dimension components whose deviation exceeds the preset support vector interval. The deviation dimension component is input into a preset fault mechanism topology diagram to lock in the coupling influencing factors; The projection length of the feature vector in the direction of the coupling influencing factor is calculated to obtain the coupling influence index.

5. The automated control method for alloy saw blade production and processing according to claim 1, characterized in that, The process of identifying the dominant anomaly source from the coupled influence indicators, and if the dominant anomaly source is uneven cutting load, involves adjusting the target rotational speed and target feed rate to determine the adjustment scheme, including: Principal component extraction is performed on the coupling influence index to determine the dominant anomaly source. If the dominant anomaly source is uneven cutting load, the real-time cutting torque and machine tool spindle stiffness matrix are obtained, and stability calculation is performed to determine the speed limit range. Within the specified speed limit range, the value with the smallest vibration amplitude is selected as the target speed. The target rotational speed is mapped onto a preset stability lobe diagram to determine the critical stable cutting depth corresponding to the target rotational speed; Based on the preset machining efficiency requirements, the target feed rate is determined within the limit range of the critical stable cutting depth; The target rotational speed and the target feed rate are combined to determine the adjustment scheme.

6. The automated control method for alloy saw blade production and processing according to claim 1, characterized in that, The process involves updating control commands based on the adjustment scheme and detecting feedback vibration signals. If the feedback vibration signals tend to stabilize, an optimization model is constructed and optimization processing is performed to determine the efficiency improvement path, including: The adjustment scheme is mapped to G-code instructions to drive the machine tool actuator and collect the feedback vibration signal when the actuator moves. The feedback vibration signal is filtered to obtain a clean signal; Extract characteristic frequencies from the pure signal and calculate the rate of change of the amplitude of the characteristic frequencies; If the amplitude change rate is less than the preset stable threshold, the corresponding adjustment scheme is determined as the steady-state parameter, and the corresponding material removal rate is recorded. An optimization model is constructed based on the steady-state parameters and the material removal rate. The optimization model is then optimized to determine the path to improve efficiency.

7. The automated control method for alloy saw blade production and processing according to claim 1, characterized in that, The process involves monitoring subsequent signals based on the efficiency improvement path and performing multi-resolution, multi-layer wavelet decomposition to determine multi-scale components. Based on the decay rate of the residual components in the multi-scale components and the downtime cost function, a downtime loss reduction threshold is determined, including: Based on the efficiency improvement path, the acquired signal is discretized and sampled to determine the digital signal; The digital signal is subjected to multi-resolution multi-layer wavelet decomposition to determine multi-scale components, residual components are identified from the multi-scale components, and the decay rate of the residual components is calculated. If the decay rate is greater than the preset convergence threshold, the material utilization rate is calculated to obtain the utilization rate. Calculate the deviation between the utilization rate and the preset benchmark utilization rate to determine the amount of reduction; The reduction is mapped to a preset downtime cost function for quantitative evaluation to determine the threshold for reducing downtime losses.

8. The automated control method for alloy saw blade production and processing according to claim 1, characterized in that, The process of extracting the iteration step size from the downtime loss reduction threshold, determining the target increment, reconstructing the transient stress distribution based on the target increment to determine the coupling variables, performing iterative separation based on the coupling variables, determining the loop conditions, and outputting a minimization control command includes: The original trajectory during the intermittent cutting process is obtained, and the original trajectory is used to extract features to determine the cutting trajectory features; The iteration step size is extracted from the downtime loss reduction threshold to determine the initial increment of the intermittent cutting process. If the peak value of the pulse load in the cutting trajectory features exceeds the preset stress limit, the initial increment is corrected to determine the target increment. Based on the target increment, the transient stress distribution is reconstructed, the compensation vector is calculated, and the compensation vector is subjected to multidimensional mapping processing to determine the coupling variables; Identify the response delay characteristics of the coupled variable; if the response delay characteristics are within a preset stable range, perform iterative separation of high-frequency and low-frequency interference to determine the loop condition. The cyclic condition is injected into the control logic to perform interference hedging within the intermittent cutting coverage area and output a minimize control command.

9. An automated control system for the production and processing of alloy saw blades, characterized in that, include: The signal acquisition module is used to acquire vibration signals during the processing of alloy saw blades, convert the vibration signals to the frequency domain, and determine abnormal frequency distributions. The symptom identification module is used to extract feature vectors from the abnormal frequency distribution. If the energy proportion of a specific frequency band in the feature vector exceeds a preset separation threshold, the corresponding feature vector is determined as an abnormal separation vector. The factor quantification module is used to calculate and decompose Euclidean distance based on the anomaly separation vector using a preset support vector machine model, and to lock the coupling influencing factors and determine the coupling influence index according to the preset fault mechanism topology map. The scheme determination module is used to identify the dominant anomaly source from the coupled influence indicators. If the dominant anomaly source is uneven cutting load, the target speed and target feed rate are adjusted to determine the adjustment scheme. The path optimization module is used to update control commands based on the adjustment scheme and detect feedback vibration signals. If the feedback vibration signals tend to stabilize, an optimization model is constructed and optimization processing is performed to determine the efficiency improvement path. The threshold evaluation module is used to monitor subsequent signals according to the efficiency improvement path and perform multi-resolution multi-layer wavelet decomposition to determine multi-scale components. Based on the decay rate of the residual components in the multi-scale components and the downtime cost function, the downtime loss reduction threshold is determined. The iterative control module is used to extract the iteration step size from the downtime loss reduction threshold, determine the target increment, reconstruct the transient stress distribution based on the target increment to determine the coupling variable, perform iterative separation according to the coupling variable, determine the loop condition, and output the minimization control command.

Images