Intelligent control method and system for high-speed milling disc milling cutter

Abstract

The application discloses a high-speed milling disc milling cutter intelligent control method and system, relates to the workpiece milling technical field, and comprises the following steps: acquiring a state signal in a cutting process; time-synchronously associating the state signal with a cutting action of the milling disc milling cutter to obtain a synchronous association signal; extracting an instantaneous impact feature corresponding to each blade cutting instant in the milling disc milling cutter according to the synchronous association signal; periodically accumulating and statistically analyzing a plurality of the instantaneous impact features to generate a health baseline; judging blade local damage according to the current vibration signal, the instantaneous impact feature and the health baseline to obtain a blade local damage judgment result; and performing hierarchical protection control according to the blade local damage judgment result. The application can combine the instantaneous impact feature and the health baseline to judge the blade local damage, so that intelligent control of the milling disc milling cutter is realized, and the accuracy and reliability are improved.

Description

Technical Field

[0001] This invention relates to the field of workpiece milling technology, and in particular to an intelligent control method and system for high-speed milling disc cutters. Background Technology

[0002] In modern manufacturing, the use of disc milling cutters is crucial for achieving high-efficiency and high-precision production. Existing systems dynamically adjust machine tool parameters by collecting and analyzing various machining data in real time, such as cutting force, vibration, and temperature, to ensure the stability and accuracy of the machining process, improve production efficiency, and extend tool life. However, in actual production environments, during high-speed milling and finishing of high-value aero-engine casings, tiny hard points inside the workpiece can cause localized micro-damage to individual disc milling cutter inserts. This damage can lead to chipping of the insert edge, loss of normal cutting ability, and consequently, subsequent inserts will bear double the cutting load, resulting in a significantly periodic and unbalanced cutting process. When a single insert experiences a momentary overload, the reflected total load current of the spindle motor is merely a negligible spike, easily drowned out by background noise. This makes it difficult for the system to identify as a serious abnormal event, hindering the adjustment of corresponding cutting parameters, resulting in low cutting control accuracy and reliability.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0004] The main objective of this invention is to propose an intelligent control method and system for high-speed milling disc cutters, which can combine instantaneous impact characteristics and health baseline to determine local damage to the cutting tool, thereby achieving intelligent control of the milling disc cutter and improving accuracy and reliability.

[0005] On one hand, embodiments of the present invention provide an intelligent control method for high-speed milling disc cutters, comprising the following steps:

[0006] Acquire state signals during the cutting process, including load signals and current vibration signals;

[0007] The state signal is time-synchronized with the cutting action of the milling disc cutter to obtain a synchronization correlation signal;

[0008] Based on the synchronous correlation signal, extract the instantaneous impact characteristics corresponding to the cutting moment of each insert in the milling disc cutter;

[0009] A health baseline is generated by periodically accumulating and statistically analyzing multiple instantaneous impact characteristics.

[0010] Based on the current vibration signal, the instantaneous impact characteristics, and the health baseline, a local damage assessment of the blade is performed to obtain the local damage assessment result of the blade.

[0011] Based on the results of the assessment of local damage to the blade, graded protection and control are implemented.

[0012] On the other hand, embodiments of the present invention provide an intelligent control system for high-speed milling disc cutters, comprising:

[0013] The information acquisition module is used to acquire status signals during the cutting process, including load signals and current vibration signals;

[0014] The time synchronization association module is used to synchronize the status signal with the cutting action of the milling disc cutter to obtain a synchronization association signal;

[0015] The feature extraction module is used to extract the instantaneous impact features corresponding to the cutting moment of each insert in the milling disc cutter based on the synchronous correlation signal;

[0016] The baseline generation module is used to periodically accumulate and statistically analyze multiple instantaneous impact characteristics to generate a healthy baseline;

[0017] The blade damage assessment module is used to assess local blade damage based on the current vibration signal, the instantaneous impact characteristics, and the health baseline, and to obtain the blade local damage assessment result.

[0018] The protection control module is used to perform graded protection control based on the judgment result of local damage to the blade.

[0019] The embodiments of this application include at least the following beneficial effects: First, the state signals during the cutting process are acquired, and the state signals are time-synchronized with the cutting action of the milling disc cutter to obtain a synchronization correlation signal. Then, based on the synchronization correlation signal, the instantaneous impact characteristics corresponding to the cutting moment of each insert in the milling disc cutter are extracted, and multiple instantaneous impact characteristics are periodically accumulated and statistically analyzed to generate a health baseline. Then, based on the current vibration signal, instantaneous impact characteristics, and health baseline, local damage to the insert is judged to obtain the local damage judgment result. Finally, based on the local damage judgment result, graded protection control is performed, thereby enabling the judgment of local damage to the insert by combining instantaneous impact characteristics and health baseline, so as to realize intelligent control of the milling disc cutter and improve accuracy and reliability.

[0020] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description and the drawings. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below.

[0022] Figure 1 This is a flowchart illustrating an intelligent control method for high-speed milling disc cutters according to an embodiment of the present invention;

[0023] Figure 2 This is a schematic diagram of the structure of an intelligent control system for a high-speed milling disc cutter according to an embodiment of the present invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments.

[0025] In modern manufacturing, particularly for high-value, complex parts, high-speed milling, especially using disc milling cutters, is crucial for achieving high-efficiency and high-precision production. This type of process is widely used in industries such as aerospace and automotive manufacturing, which have stringent requirements for material removal rates and final surface quality. To address the complexities that may arise during high-speed milling, such as cutting force fluctuations, tool wear, and potential vibrations, intelligent control systems are typically employed. These intelligent systems dynamically adjust machine tool parameters, such as spindle speed, feed rate, and depth of cut, by collecting and analyzing various machining data in real time, including cutting forces, vibrations, and temperature. Their primary goal is to ensure the stability and accuracy of the machining process while simultaneously improving production efficiency and extending tool life.

[0026] In real-world production environments, even machine tools equipped with advanced intelligent control systems can encounter unforeseen challenges. For example, in a large equipment manufacturing workshop, a five-axis CNC machining center is performing a critical task: final surface finishing of a large, nearly finished aero-engine casing. This casing, cast from a high-temperature alloy, is extremely valuable, and any machining error could render the entire part unusable. To ensure machining efficiency and surface quality, the machine tool uses a 200mm diameter high-speed disc milling cutter with 12 evenly spaced carbide-coated inserts on the cutter head. The machine tool itself integrates an intelligent control system, initially designed to automatically adjust the feed rate and spindle speed by monitoring the spindle motor's load current and vibration sensor signals mounted on the spindle head in real time, thereby maintaining the cutting process in a stable and efficient state. Under ideal machining conditions, when the tool cuts into the workpiece, the system will fine-tune the feed according to the smooth increase of the load to ensure that the cutting force borne by each insert is within the preset ideal range. If the system detects that the vibration of certain frequencies begins to increase, it will judge it as an early sign of machining chatter and actively fine-tune the spindle speed to avoid the unstable cutting resonance area.

[0027] However, despite undergoing rigorous flaw detection, a tiny but extremely hard carbide inclusion remained in a non-critical location within the high-temperature alloy casing during the casting process. This hard spot was located and sized in a blind spot of the inspection, and therefore went undetected before machining. When the high-speed rotating disc milling cutter swept across this area, one of the inserts made a momentary, violent impact with this hard spot. The impact was extremely brief, possibly only a few milliseconds. For the entire system, which involves 12 inserts cutting together, this momentary overload of a single insert resulted in a negligible spike in the total load current of the spindle motor, almost drowned out by background noise. Therefore, the intelligent control system, based on its total load current judgment logic, did not recognize it as a serious anomaly and failed to take any emergency control actions. It was difficult to adjust the corresponding cutting parameters, resulting in low accuracy and reliability of the cutting control.

[0028] When performing high-speed milling and finishing on high-value aero-engine casings, the control system needs to overcome the diagnostic blind spot of weak instantaneous impacts in the global signal, which may be caused by tiny hard points inside the workpiece. It also needs to accurately distinguish between periodic low-frequency vibrations caused by tool damage and conventional machining chatter, so as to achieve early and accurate diagnosis of local tool damage and trigger appropriate protective control responses to avoid the failure from escalating to the scrapping of high-value workpieces and equipment damage.

[0029] In view of this, this application acquires the state signals during the cutting process and synchronously correlates them with the cutting action of the milling disc cutter, thereby accurately capturing the instantaneous impact characteristics of each insert at the moment of cutting. By periodically accumulating and statistically analyzing these instantaneous impact characteristics, a health baseline is generated, providing a reliable reference for subsequent assessment of local insert damage. Finally, based on the assessment results of local insert damage, graded protection control is implemented, effectively avoiding the limitations of traditional systems in identifying local insert damage, significantly improving the stability and safety of the high-speed milling process, and reducing the risk of scrapping high-value workpieces.

[0030] The embodiments of this application will be explained in detail below with reference to the accompanying drawings:

[0031] Figure 1 This is an optional flowchart of an intelligent control method for high-speed milling disc cutters provided in an embodiment of this application. Figure 1 The method may include, but is not limited to, steps S101 to S106.

[0032] Step S101: Acquire the status signals during the cutting process. The status signals include the load signal and the current vibration signal.

[0033] Step S102: Synchronize the status signal with the cutting action of the milling disc cutter in time to obtain a synchronization signal;

[0034] Step S103: Based on the synchronous correlation signal, extract the instantaneous impact characteristics corresponding to the cutting moment of each insert in the milling disc cutter;

[0035] Step S104: Perform periodic accumulation and statistical analysis on multiple instantaneous impact characteristics to generate a health baseline;

[0036] Step S105: Based on the current vibration signal, instantaneous impact characteristics and health baseline, determine the local damage of the blade and obtain the local damage determination result of the blade.

[0037] Step S106: Based on the results of the local damage assessment of the blade, implement graded protection and control.

[0038] Steps S101 to S106 shown in the embodiments of this application can combine instantaneous impact characteristics and health baseline to determine local damage to the cutting tool, thereby realizing intelligent control of the milling disc cutter and improving accuracy and reliability.

[0039] In some embodiments, steps S101-S106 can first acquire state signals during the cutting process, whereby the state signals include load signals and current vibration signals. The load current of the spindle motor can be monitored in real time by installing a current sensor on the spindle of the milling cutter, and this load signal can be used as the load signal. Simultaneously, a high-sensitivity accelerometer can be installed on the machine tool spindle box or workpiece fixture to collect the current vibration signal. These sensors should have sufficient sampling frequency and accuracy to ensure that transient changes during the cutting process can be captured. It is understood that the load signal refers to the spindle motor current, the force signal output by the cutting force sensor, etc., used to reflect the cutting load borne by the cutting tool during the cutting process. The current vibration signal refers to the vibration data collected by the accelerometer or acoustic emission sensor, used to reflect the vibration characteristics of the tool or workpiece during the cutting process. Accurate acquisition of these signals is the basis for subsequent analysis and judgment.

[0040] The status signal is then time-synchronized with the cutting action of the disc milling cutter to obtain a synchronization signal. The spindle's angular position can be precisely acquired by installing a rotary encoder on the spindle, thus determining the precise time points when each insert enters and exits the cutting zone. Time synchronization is achieved by aligning the time series of the load signal and the current vibration signal with these time points. Alternatively, the toolpath and spindle speed information provided by the machine tool's CNC system can be used, combined with tool geometry parameters, to calculate the theoretical cutting time window for each insert, and the status signal can be correlated with these windows. In essence, the cutting action of a disc milling cutter refers to the periodic movement of each insert sequentially entering and exiting the cutting zone during the machining process.

[0041] Then, based on the synchronous correlation signal, the instantaneous impact characteristics corresponding to the cutting moment of each insert in the milling disc cutter are extracted. For each insert's cutting time window, features such as peak value, rise time steepness, and energy can be extracted from the synchronously correlated load signal as instantaneous impact characteristics. For vibration signals, instantaneous amplitude, impact factor, or kurtosis within a specific frequency range can be extracted as instantaneous impact characteristics. These characteristics can effectively characterize the impact intensity and characteristics experienced by the insert during the cutting moment. In essence, instantaneous impact characteristics refer to the transient impact response generated during the contact, cutting, and separation processes between the insert and the workpiece material at each cutting moment. These impact characteristics contain key information about the insert's health condition, such as whether the insert has chipped or worn.

[0042] A healthy baseline is generated by periodically accumulating and statistically analyzing multiple instantaneous impact characteristics. Statistical analysis can be performed on the instantaneous impact characteristics corresponding to each insert, calculating its mean, standard deviation, maximum, minimum, and other statistical quantities. Through long-term accumulation and statistical analysis, a distribution model of the instantaneous impact characteristics of each insert under healthy conditions can be established, i.e., the healthy baseline. In essence, the healthy baseline refers to a reference standard established through long-term accumulation and statistical analysis of instantaneous impact characteristics when the milling disc cutter is in a normal, healthy cutting state. The healthy baseline reflects the inherent fluctuation range and statistical regularity of instantaneous impact characteristics under normal cutting conditions, providing a benchmark for judging whether an insert has experienced abnormalities.

[0043] Based on the current vibration signal, instantaneous impact characteristics, and healthy baseline, local damage to the blade is assessed, yielding the assessment result. The instantaneous impact characteristics of each blade are compared with the corresponding healthy baseline. If the instantaneous impact characteristics of a blade significantly deviate from its healthy baseline, and a specific frequency component related to blade damage appears in the current vibration signal, it can be preliminarily determined that the blade may have local damage.

[0044] Finally, based on the assessment of localized damage to the cutting tool, tiered protection control is implemented. If the assessment indicates minor damage, the system can issue a warning and advise the operator to inspect the tool before machining the next workpiece. If the assessment indicates severe damage, the system can immediately reduce the feed rate, spindle speed, or even pause machining to prevent further damage and protect both the workpiece and the machine tool.

[0045] This embodiment achieves precise capture of the instantaneous impact characteristics of each cutting tool during the cutting process through refined data acquisition and analysis, and establishes a dynamically updated health baseline based on this. Specifically, firstly, by acquiring the load signal and current vibration signal during the cutting process and synchronously correlating them with the cutting action of the milling disc cutter, the precise correspondence between the data and the specific cutting event of the cutting tool is ensured, thus overcoming the limitations of traditional methods such as signal aliasing and difficulty in locating the damage source. Secondly, by extracting the instantaneous impact characteristics of each cutting tool during the cutting moment based on the synchronously correlated signal, it can more sensitively reflect the changes in the microscopic state when the cutting tool is in contact with the workpiece material, and even minor chipping or wear can be effectively identified. Furthermore, by periodically accumulating and statistically analyzing multiple instantaneous impact characteristics, a health baseline is generated, providing a dynamic and adaptive reference standard for cutting tool damage assessment. The health baseline can better adapt to different processing conditions and material properties, reducing false alarms and missed alarms. Finally, by combining the current vibration signal, instantaneous impact characteristics, and health baseline, local damage to the cutting tool is assessed, and graded protection control is implemented based on the assessment results, realizing a multi-level response from early warning to emergency shutdown. This tiered protection mechanism effectively prevents the scrapping of high-value workpieces and equipment damage, significantly improving the reliability and safety of the high-speed milling process.

[0046] Through the above technical solutions, this embodiment constructs a more precise and intelligent high-speed milling disc cutter control method by introducing techniques such as time synchronization correlation, instantaneous impact feature extraction, health baseline generation, and hierarchical protection control. Compared with existing technologies, this embodiment can identify local damage to the cutting tool earlier and more accurately, and take appropriate protective measures, thereby significantly improving processing efficiency, reducing production costs, and extending the service life of cutting tools and machine tools.

[0047] In some embodiments, in step S105, the local damage assessment of the blade is performed based on the current vibration signal, instantaneous impact characteristics, and healthy baseline to obtain the local damage assessment result of the blade, which may include, but is not limited to, the following steps:

[0048] Calculate the first degree of deviation between the instantaneous impact characteristics and the healthy baseline;

[0049] If the first deviation degree is greater than the first deviation threshold, then determine whether there is a periodic correlation between the first deviation degree and the spindle speed;

[0050] If the degree of the first deviation is periodically correlated with the spindle speed, then frequency analysis is performed on the current vibration signal to obtain the frequency analysis results;

[0051] Based on the frequency analysis results, the target vibrations were determined, including conventional machining vibrations and periodic low-frequency vibrations caused by tool damage.

[0052] Based on the degree of deviation and the target vibration, a judgment result on the local damage of the blade is generated.

[0053] In some embodiments, the initial deviation of the instantaneous impact characteristics from the health baseline can be calculated first. The blade condition can be assessed by quantifying the difference between the instantaneous impact characteristics and the pre-established health baseline. For example, the initial deviation can be calculated using statistical distance (such as Euclidean distance, Mahalanobis distance) or percentage deviation, with the aim of initially identifying whether there are any abnormalities in the blade's cutting condition.

[0054] If the first deviation exceeds a first deviation threshold, it is determined whether there is a periodic correlation between the first deviation and the spindle speed. The first deviation threshold can be set based on historical data, statistical process control (SPC) methods, or expert experience to distinguish between normal fluctuations and potential anomalies. The purpose of determining a periodic correlation is to identify whether the anomaly is directly related to the rotational motion of the cutting tool, because localized damage to the cutting tool typically generates periodic impacts during each cutting operation, the frequency of which is closely related to the spindle speed and the number of cutting tools. For example, the existence of a periodic correlation can be determined by performing spectral analysis on the first deviation signal and comparing it with the frequency components of the spindle speed, or by using cross-correlation analysis.

[0055] If the degree of the first deviation is periodically correlated with the spindle speed, then frequency analysis is performed on the current vibration signal to obtain the frequency analysis results. Frequency analysis can employ methods such as Fast Fourier Transform (FFT) and wavelet transform to convert the time-domain vibration signal into a frequency-domain signal in order to identify the vibration energy of different frequency components. The purpose is to extract specific frequency features related to blade damage from the vibration signal.

[0056] Then, based on the frequency analysis results, the target vibrations are determined. These target vibrations include conventional machining vibrations and periodic low-frequency vibrations caused by tool damage. Conventional machining vibrations refer to background vibrations caused by machine tool structure, transmission system, workpiece material inhomogeneity, etc., and their frequency distribution may be relatively wide. Periodic low-frequency vibrations caused by tool damage, such as tool chipping and wear, will generate periodic impacts during tool cutting, and will appear in the vibration spectrum as specific low-frequency peaks related to the tool passing frequency (spindle speed multiplied by the number of tools) and its harmonics. By analyzing the frequency analysis results, these vibration components from different sources can be identified and distinguished.

[0057] Based on the degree of deviation and the target vibration, a local damage assessment result for the blade is generated. This result can be a binary judgment (e.g., "damage present / absent") or a damage level or type (e.g., "minor wear," "chipping"). The purpose is to comprehensively consider the degree of deviation and vibration characteristics to provide a complete and accurate diagnosis of local blade damage.

[0058] To illustrate this technical solution more clearly, a specific example is used below. Suppose that during high-speed milling, the system detects a significant deviation between the instantaneous impact characteristics of a cutting insert and the healthy baseline. First, the degree of deviation of this instantaneous impact characteristic from the healthy baseline is calculated. For example, it is found that the deviation reaches 15% of the average value of the healthy baseline, exceeding a preset first deviation threshold (e.g., 10%). Subsequently, the system further analyzes whether this deviation is periodically related to the spindle speed. By performing spectral analysis on the deviation signal, a significant frequency component corresponding to the insert passage frequency (spindle speed multiplied by the number of inserts) is found, indicating that the deviation is periodic and related to the insert's rotational motion. Therefore, the system performs frequency analysis on the current vibration signal, obtaining a detailed spectrum. In the spectrum, in addition to the conventional machine tool background vibration frequency, a periodic low-frequency vibration peak is clearly identified near the insert passage frequency, with an amplitude significantly higher than that during normal machining. This is determined to be periodic low-frequency vibration caused by insert damage. Finally, combining the initial deviation level with the identified periodic low-frequency vibrations caused by blade damage, the system generates a localized blade damage assessment result, clearly indicating the presence of localized damage. Through this series of refined assessment steps, localized blade damage can be accurately identified, avoiding misjudgments that might occur based on a single deviation indicator, thus providing a reliable decision-making basis for subsequent intelligent control.

[0059] Through the above technical solution, this embodiment can significantly improve the accuracy and reliability of judging local damage to high-speed milling disc cutter inserts. This embodiment can not only promptly detect abnormal insert conditions, but also effectively distinguish genuine local insert damage from similar signals caused by other factors by introducing periodic correlation judgment and frequency analysis, thereby avoiding misjudgments and unnecessary downtime or tool replacement. This makes the diagnosis of insert damage more refined, providing a more solid and accurate basis for subsequent graded protection control, thereby extending tool life and improving machining efficiency and product quality.

[0060] In some embodiments, step S104 involves periodically accumulating and statistically analyzing multiple instantaneous impact characteristics to generate a health baseline, which may include, but is not limited to, the following steps:

[0061] Step S201: Identify the current group of cutting blades within the current cutting area;

[0062] Step S202: Calculate the first statistical feature based on the instantaneous impact characteristics of each current blade in the current blade group. The first statistical feature includes the first mean and the first standard deviation.

[0063] Step S203: Identify the minimum fluctuation range based on the instantaneous impact characteristics of each current blade in the current blade group;

[0064] Step S204: Determine the health fluctuation range based on the minimum fluctuation range;

[0065] Step S205: If the first statistical characteristic meets the health fluctuation range, then generate a health baseline based on the first statistical characteristic and the health fluctuation range;

[0066] Step S206: If the first statistical characteristic does not meet the healthy fluctuation range, then perform baseline adaptive adjustment processing based on the current blade population to generate a healthy baseline.

[0067] In some embodiments, factors such as cutting conditions, workpiece material properties, and tool wear conditions may dynamically change, and simple periodic accumulation and statistical analysis may not accurately capture these changes, resulting in an insufficiently robust health baseline or an inability to reflect the true health status in a timely manner. For example, if the initial baseline is established on atypical or highly fluctuating data, or if, during long-term machining, minor changes in the environment or materials cause a slow drift in the overall impact characteristics, damage assessment based on a fixed baseline may result in false alarms or missed alarms. This could lead to inaccurate assessment of tool damage by the system, thereby affecting the effectiveness of graded protection control, and may even shorten tool life or affect machining quality.

[0068] To achieve this, the current group of cutting inserts within the current cutting area can be identified first. By analyzing the milling cutter's rotational speed, the number of inserts, and the time points when each insert enters and leaves the cutting area during the cutting process, the set of inserts currently cutting within a specific time window can be determined. For example, real-time spindle rotation information can be obtained using an encoder or speed sensor, combined with the physical distribution of the inserts on the milling cutter, to accurately identify the inserts currently in contact with the workpiece. The aim is to focus the analysis on the inserts actually involved in the cutting, thereby improving the relevance and accuracy of the data analysis.

[0069] Then, based on the instantaneous impact characteristics of each current blade in the current blade group, a first statistical characteristic is calculated. This first statistical characteristic includes a first mean and a first standard deviation. The first mean characterizes the central tendency of the instantaneous impact characteristics of the current blade group, reflecting the average impact level of the group during the cutting process. The first standard deviation quantifies the dispersion of the instantaneous impact characteristics, reflecting the volatility or consistency of the group's impact characteristics. These statistical characteristics are the fundamental data for assessing the health status of the blades.

[0070] Next, based on the instantaneous impact characteristics of each current insert in the current insert group, the minimum fluctuation range is identified. This can be achieved by analyzing the instantaneous impact characteristic data of the current insert group over multiple cutting cycles to find the data segment or set of data points with the smallest fluctuations and the greatest stability. This can be done through sliding window analysis, clustering algorithms, or methods based on statistical significance tests. The aim is to eliminate the influence of abnormal impacts or transient noise and focus on stable data that reflects the normal cutting behavior of the inserts.

[0071] Based on the minimum fluctuation range, the healthy fluctuation range is determined. A reasonable upper and lower limit can be set based on the identified minimum fluctuation range to define the normal range of variation in the instantaneous impact characteristics of the blade in a healthy state. This range can be determined by adding or subtracting a preset multiple of the standard deviation from the average value of the minimum fluctuation range, or by training with historical health data. The purpose is to provide a clear reference boundary for subsequent assessment of local blade damage.

[0072] If the first statistical characteristic meets the healthy fluctuation range, then a healthy baseline is generated based on the first statistical characteristic and the healthy fluctuation range. This indicates that the cutting state of the current blade group is in a normal and stable healthy state. At this time, the currently calculated first statistical characteristic (such as the first mean and the first standard deviation) and the determined healthy fluctuation range can be directly used as the current healthy baseline.

[0073] If the first statistical characteristic does not meet the health fluctuation range, a baseline adaptive adjustment is performed based on the current blade population to generate a healthy baseline. This indicates that the cutting state of the current blade population may have changed and no longer fully conforms to the preset health fluctuation range. In this case, the system will not simply classify it as damage, but will activate the adaptive adjustment mechanism to correct or update the health baseline based on the actual data of the current blade population, so that it can better reflect the current actual health status.

[0074] To illustrate this technical solution more clearly, a specific example is used below. Assume that during a high-speed milling operation, slight hardness differences exist between batches of workpiece material, or the machine tool experiences minor thermal deformation after prolonged operation, causing a slow, uniform drift in the instantaneous impact characteristics of the cutting insert. This embodiment first identifies the current group of cutting inserts within the current cutting area and calculates its first statistical characteristics, including a first average and a first standard deviation. Next, it identifies the minimum fluctuation range and determines the healthy fluctuation range. If the first statistical characteristics deviate from the initially set healthy fluctuation range, but this deviation is uniform and slow, and through subsequent baseline adaptive adjustment, the system can identify that this is not localized insert damage, but rather a normal drift in the overall cutting state. At this point, the system adaptively adjusts the healthy baseline based on the data of the current group of cutting inserts, adapting it to the new cutting state, thereby avoiding misjudgment and continuing to accurately monitor the localized damage of the cutting inserts.

[0075] Through the above technical solution, this embodiment significantly improves the accuracy of judging local damage to the cutting tool and reduces the risk of false alarms and missed alarms. Especially when there are slight changes in cutting conditions or workpiece materials, the adaptive adjustment mechanism can ensure that the baseline always reflects the true health status of the tool, thereby providing a more reliable basis for subsequent graded protection control, extending tool life, and improving machining efficiency and safety.

[0076] In some embodiments, in step S205, baseline adaptive adjustment is performed based on the current blade group to generate a healthy baseline, which may include, but is not limited to, the following steps:

[0077] Step S301: Based on the current blade group, identify the neighboring blade groups within the immediate cutting area;

[0078] Step S302: Calculate the second statistical feature based on the instantaneous impact characteristics of each neighboring blade in the neighboring blade group. The second statistical feature includes the second mean and the second standard deviation.

[0079] Step S303: Update the health fluctuation range based on the second statistical characteristic;

[0080] Step S304: Generate a health baseline based on the second statistical feature and the updated health fluctuation range.

[0081] In some embodiments, since baseline adjustment is based solely on the current group of cutting inserts within the current cutting area, it may not adequately address overall drift caused by slow changes in the cutting environment or workpiece material properties. Alternatively, when the current group of cutting inserts is already in an unhealthy state but has not yet been identified as damaged, adaptive adjustments based on it may introduce biases, thereby affecting the accuracy of the healthy baseline and the reliability of subsequent damage assessments.

[0082] To this end, we can first identify neighboring insert groups within the cutting area based on the current insert group. This can be done by selecting a set of inserts that are temporally or spatially adjacent to the current cutting area and are considered to be in a relatively stable or healthy state, outside of the currently cutting insert group. This neighboring insert group can be inserts in areas where cutting has already been completed on the current workpiece, or inserts that have been cut on the same batch of workpieces and are considered to be undamaged. The purpose is to provide a more stable and reliable reference base to avoid potential anomalies in the current insert group interfering with baseline adjustment.

[0083] Then, based on the instantaneous impact characteristics of each neighboring blade in the neighboring blade group, a second statistical characteristic is calculated, which includes a second mean and a second standard deviation. Statistical analysis can be performed on the instantaneous impact characteristics of each neighboring blade in the identified neighboring blade group. These statistical characteristics provide a quantitative basis for subsequent updates to the health fluctuation range.

[0084] Then, based on the second statistical characteristic, the health fluctuation range is updated. The original health fluctuation range can be corrected by utilizing more stable cutting state information reflected by neighboring blade groups. For example, the upper and lower limits of the health fluctuation range can be recalculated or adjusted based on the second mean and second standard deviation to more accurately reflect the current actual cutting health.

[0085] Finally, a health baseline is generated based on the second statistical characteristic and the updated health fluctuation range. After obtaining more reliable statistical characteristics and the updated health fluctuation range, this information can be combined to construct or revise the health baseline. This health baseline will be more robust and can more accurately reflect the instantaneous impact characteristic distribution of the milling disc cutter under normal operating conditions, providing a solid foundation for subsequent assessment of localized insert damage.

[0086] To illustrate this technical solution more clearly, a specific example is used below. Suppose that during the machining of a batch of workpieces, a high-speed milling disc cutter, due to minor batch variations in workpiece material or the slow accumulation of machine tool thermal deformation, causes the first statistical characteristic of the instantaneous impact feature of the cutting edge in the current cutting area to slightly deviate from the initially set healthy fluctuation range, but has not yet reached the damage threshold. If baseline adjustments are made solely based on the current group of cutting edges, this overall drift might be misjudged as an early sign of edge damage. Conversely, if the current group of cutting edges is already slightly worn, the baseline adjusted based on it might mask the true damage.

[0087] At this point, the system identifies neighboring blade groups within the immediate cutting area based on the current blade group. For example, blades that have already completed cutting on the current workpiece and have been verified as healthy by historical data, or blades from the same batch of workpieces but at different cutting stages and considered undamaged, can be selected as neighboring blade groups. Subsequently, the system performs statistical analysis on the instantaneous impact characteristics of each neighboring blade within these groups, calculating a second statistical characteristic, including a second mean and a second standard deviation. Based on these second statistical characteristics, the system updates the health fluctuation range. For example, if the second mean shows a slight increase in the overall impact level, the upper and lower limits of the health fluctuation range will shift upwards accordingly to accommodate this normal process drift. Finally, a new health baseline is generated based on the updated health fluctuation range and the second statistical characteristic. In this way, the health baseline can more accurately reflect the current actual, healthy cutting state, effectively distinguishing between normal process drift and genuine localized blade damage, thereby avoiding misjudgment and ensuring the accuracy of blade damage assessment.

[0088] By incorporating a group of cutting inserts within the adjacent cutting area as a reference and performing statistical analysis based on their instantaneous impact characteristics, this embodiment obtains more stable and reliable statistical features. This allows for updating the health fluctuation range, resulting in a more robust and accurate health baseline. This significantly enhances the health baseline's ability to reflect actual cutting conditions, reduces the risk of misjudging or omitting localized insert damage, and thus improves the accuracy and reliability of intelligent control for high-speed milling disc cutters.

[0089] In some embodiments, in step S304, generating a health baseline based on the second statistical characteristic and the updated health fluctuation range may include, but is not limited to, the following steps:

[0090] Step S401: Calculate the trend of feature change based on the first statistical feature and the second statistical feature;

[0091] Step S402: Calculate the rate of change of the features based on the trend of feature changes;

[0092] Step S403: Based on the characteristic change rate, perform parameter drift judgment to obtain the parameter drift judgment result;

[0093] Step S404: Based on the parameter drift judgment result, the first statistical feature and the second statistical feature are weighted and fused to obtain the target statistical feature;

[0094] Step S405: Generate a health baseline based on the target statistical characteristics and the updated health fluctuation range.

[0095] In some embodiments, changes in instantaneous impact characteristics may not be solely due to localized damage to the cutting tool, but may also be influenced by various factors such as uniform wear of the tool as a whole, aging of machine tool components, or changes in workpiece material properties. If these different types of parameter drift are not effectively distinguished, directly using the second statistical characteristics of neighboring cutting tool groups to generate a health baseline may result in inaccurate baseline adjustment, thereby affecting the accuracy of the assessment of localized cutting tool damage.

[0096] Therefore, the trend of feature change can be calculated based on the first and second statistical characteristics. The difference or ratio between the first statistical characteristic corresponding to the current group of inserts within the current cutting region and the second statistical characteristic corresponding to the adjacent group of inserts within the immediate cutting region can be quantified. For example, the difference, ratio, or relative rate of change between the first and second statistical characteristics can be calculated to reflect the direction and magnitude of the evolution of instantaneous impact characteristics among different groups of inserts. The purpose is to preliminarily assess the overall change of instantaneous impact characteristics, providing basic data for subsequent in-depth analysis.

[0097] Then, based on the characteristic change trend, the characteristic change rate is calculated. Based on obtaining the characteristic change trend, the rate of change of this trend with time or the cutting area can be further analyzed. For example, the characteristic change rate can be obtained by performing derivative, difference, or regression analysis on the characteristic change trends of multiple consecutive cutting cycles or different cutting areas. The purpose is to identify the dynamic characteristics of instantaneous impact characteristic changes, such as whether the change is slow and uniform or rapid and sudden, providing crucial information for subsequent parameter drift judgment.

[0098] Next, based on the rate of change of the characteristics, parameter drift is determined, yielding the parameter drift assessment result. The calculated rate of change of characteristics can be used to distinguish the specific causes of the instantaneous impact characteristic changes. For example, parameter drift may include uniform wear of the tool as a whole, non-uniform drift caused by aging of machine tool components, or global drift. This assessment process aims to identify whether the overall change in instantaneous impact characteristics is caused by normal tool wear, machine tool system performance degradation, or other global factors, thus providing precise guidance for subsequent baseline adjustments.

[0099] Based on the parameter drift assessment results, the first and second statistical features are weighted and fused to obtain the target statistical feature. Different weights can be assigned to the first and second statistical features according to the parameter drift assessment results to generate a more representative and robust target statistical feature. For example, if the assessment result indicates uniform wear of the tool as a whole, the weight of the second statistical feature can be appropriately increased, as it better reflects the current overall wear state; if the assessment indicates non-uniform drift caused by aging of machine tool components, a more complex weighting strategy, or even the introduction of a compensation model, may be required. The aim is to comprehensively consider the statistical information of different tool groups and intelligently adjust according to the drift type to eliminate or reduce the impact of non-damaging factors on the health baseline.

[0100] Finally, a health baseline is generated based on the target statistical characteristics and the updated health fluctuation range. The weighted fusion-processed target statistical characteristics can be combined with the updated health fluctuation range to determine the health baseline used for assessing localized blade damage. This health baseline will more accurately reflect the normal impact level under the current cutting condition and effectively eliminate interference caused by non-damaging factors.

[0101] To illustrate this technical solution more clearly, a specific example is used below. Suppose that during high-speed milling, the system detects that the first statistical characteristic corresponding to the current group of cutting inserts exceeds a preset health fluctuation range. At this point, the system identifies neighboring groups of cutting inserts within the immediate cutting area and calculates their second statistical characteristic. If the baseline is adjusted directly using the second statistical characteristic according to the basic solution, it may be impossible to distinguish whether the inserts are experiencing localized damage or the entire tool is undergoing uniform wear. In this case, the system calculates the characteristic change trend between the first and second statistical characteristics, for example, calculating the difference between their average values. Then, by analyzing these trends over multiple consecutive cutting cycles or different cutting areas, the characteristic change rate is calculated. For example, if the average value is found to be increasing slowly and uniformly, the characteristic change rate will show a stable positive value. Subsequently, the system performs parameter drift judgment based on this characteristic change rate. For example, if the characteristic change rate is less than a preset wear change threshold, and spectral analysis of historical vibration signals from target components such as machine tool spindle bearings or guideways reveals no localized or non-unidirectional slow drift, then the tool is judged to be experiencing uniform overall wear.

[0102] Once the system determines that the tool is experiencing uniform wear, it weights and fuses the first and second statistical features based on this assessment. For example, the second statistical feature can be assigned a higher weight (e.g., 0.7), while the first statistical feature can be assigned a lower weight (e.g., 0.3), because the second statistical feature better represents the overall wear state. Through this weighted fusion, the resulting target statistical feature more accurately reflects the normal impact level of the tool under uniform wear. Finally, the system generates a new health baseline based on this refined target statistical feature and the updated health fluctuation range. This approach effectively avoids misjudging uniform wear as localized damage, thereby improving the accuracy of tool damage assessment and the system's intelligent control level.

[0103] Through the above technical solution, this embodiment achieves intelligent identification and differentiation of different types of drift by analyzing the trend and rate of feature changes and combining parameter drift judgment. Therefore, when generating a health baseline, the first and second statistical features can be weighted and fused according to the specific drift type, thereby generating a more accurate and robust target statistical feature. This significantly improves the accuracy and adaptability of the health baseline, making the judgment of local tool damage more reliable, effectively reducing false alarm and false negative rates, and improving the overall performance and reliability of the intelligent control system for high-speed milling disc cutters.

[0104] In some embodiments, step S403, based on the characteristic change rate, performs parameter drift determination to obtain a parameter drift determination result, which may include, but is not limited to, the following steps:

[0105] Acquire historical vibration signals of the target component, which may include machine tool spindle bearings or guideways;

[0106] Spectral analysis of historical vibration signals yields the initial vibration frequency and initial amplitude.

[0107] Construct an initial vibration characteristic spectrum based on the initial vibration frequency and initial amplitude;

[0108] Monitor the current vibration signal of the target component;

[0109] Perform spectral analysis on the current vibration signal to obtain the current vibration frequency and current amplitude;

[0110] Construct the current vibration characteristic spectrum based on the current vibration frequency and current amplitude;

[0111] The initial vibration feature spectrum and the current vibration feature spectrum are compared to obtain the comparison results;

[0112] If the comparison result shows that there is no localized and non-unidirectional slow drift, and the characteristic change rate is less than the preset wear change threshold, then the parameter drift judgment result is determined to be uniform wear of the tool as a whole.

[0113] If the comparison result shows that there is localized and non-uniform slow drift, and the characteristic change rate is greater than the preset wear change threshold, then the parameter drift judgment result is determined to be non-uniform drift caused by the aging of machine tool components.

[0114] In some embodiments, parameter drift can be caused by a variety of factors, such as uniform wear of the tool or aging of critical machine tool components (such as spindle bearings or guideways). Failure to accurately distinguish these different sources of drift may lead to misjudgment of the tool's health status, thereby affecting the effectiveness of subsequent protection and control strategies.

[0115] To achieve this, historical vibration signals of the target component can be acquired first. The target component refers to a critical machine tool structure that significantly affects the vibration characteristics of the milling process, such as the machine tool spindle bearing or guideway. Acquiring these historical vibration signals aims to establish a baseline state for subsequent comparison with the current state. These historical vibration signals are typically collected when the machine tool is in a healthy and stable operating state. Spectral analysis is performed on the historical vibration signals to obtain the initial vibration frequency and initial amplitude. Based on the initial vibration frequency and initial amplitude, an initial vibration characteristic spectrum is constructed. This spectrum characterizes the inherent vibration modes of the target component under normal operating conditions.

[0116] Then, monitor the current vibration signal of the target component. During the actual cutting process of the milling disc cutter, continuously monitor the current vibration signal of the target component. Perform spectrum analysis on the current vibration signal to obtain the current vibration frequency and amplitude. Similar to historical vibration signals, the current vibration signal also needs to undergo spectrum analysis to obtain the current vibration frequency and amplitude, and construct a current vibration characteristic spectrum based on the current vibration frequency and amplitude.

[0117] The initial vibration characteristic spectrum is then compared with the current vibration characteristic spectrum to obtain the comparison results, aiming to identify changes in vibration modes, especially the presence of localized and non-unidirectional slow drifts. Localized drifts may indicate anomalies in specific frequencies or regions, while non-unidirectional drifts mean that the changes in vibration modes are neither uniform nor global.

[0118] If the comparison results show no localized or non-unidirectional slow drift, and the characteristic change rate is less than the preset wear change threshold, it indicates that the entire system (including the tool and machine tool) is in a relatively stable state, and the parameter drift is mainly attributed to the overall uniform wear of the tool. The parameter drift judgment result can be determined as overall uniform wear of the tool. In this case, the adjustment of the healthy baseline should focus on compensating for this uniform wear.

[0119] If the comparison results show localized and non-uniform slow drift, and the characteristic change rate is greater than the preset wear change threshold, it indicates that the parameter drift is not simply caused by tool wear, but may be due to aging or localized damage to machine tool components (such as spindle bearings or guideways). The parameter drift can be determined as non-uniform drift caused by the aging of machine tool components. In this case, adjusting the health baseline needs to consider the impact of the machine tool component condition, and may even trigger machine tool maintenance alarms.

[0120] To illustrate this technical solution more clearly, a specific example is used below. Assume that before a high-speed milling machine tool is put into use, comprehensive vibration signal acquisition and spectrum analysis are performed on its spindle bearings and guideways, constructing an initial vibration characteristic spectrum. During subsequent long-term machining operations, the system continuously monitors the current vibration signals of these target components. For example, in a certain milling task, the system detects that the periodic accumulation of instantaneous impact characteristics and the statistical analysis results (first statistical characteristic) do not meet the healthy fluctuation range, triggering baseline adaptive adjustment processing and calculating the characteristic change rate. When judging parameter drift, the system compares the current vibration characteristic spectrum of the spindle bearing with the initial vibration characteristic spectrum.

[0121] If the comparison results show that the overall amplitude of the spindle bearing vibration spectrum has increased slightly, but no new local peaks or non-uniform frequency drifts have appeared, and the characteristic change rate (e.g., the average change rate of instantaneous impact characteristics) is less than the preset wear change threshold, the system will determine that the parameter drift is due to the overall uniform wear of the tool. In this case, the health baseline will be adjusted according to the uniform wear trend of the tool to reflect the normal aging process of the tool. If the comparison results show that the spindle bearing vibration spectrum has a new local high amplitude peak in a specific frequency region, and this peak has a slow, non-uniform drift compared to the initial spectrum, while the characteristic change rate is greater than the preset wear change threshold, the system will determine that the parameter drift is not simply caused by tool wear, but by non-uniform drift caused by local aging or damage to the spindle bearing. In this case, in addition to adjusting the health baseline, the system will also issue a machine tool maintenance alarm to prompt the operator to check the spindle bearing, thereby avoiding a decrease in machining quality or equipment damage due to machine tool component failure. Through the above two methods of judgment, this embodiment can accurately identify the cause of parameter drift, thereby guiding more reasonable health baseline adjustments and more effective equipment maintenance strategies.

[0122] Through the above technical solution, this embodiment can effectively distinguish between different sources of parameter drift during high-speed milling disc cutter cutting, namely, uniform tool wear and non-uniform drift caused by machine tool component aging. This refined judgment avoids misjudging abnormalities caused by machine tool component aging as tool wear, thereby improving the accuracy of health baseline generation. Therefore, it not only enables more accurate assessment of the actual health status of the tool, but also provides early warning for preventative machine tool maintenance, extends machine tool life, and ensures the stability and machining quality of the milling process.

[0123] In some embodiments, step S403, based on the characteristic change rate, performs parameter drift determination to obtain a parameter drift determination result, which may include, but is not limited to, the following steps:

[0124] Step S501: Before machining the workpiece, perform a multi-cycle cutting test on the milling disc cutter to obtain initial impact characteristic data;

[0125] Step S502: Determine the initial fluctuation range based on the initial impact characteristic data;

[0126] Step S503: Calculate the drift trend based on the initial impact characteristic data;

[0127] Step S504: Calculate the second degree of deviation based on the characteristic change rate and the initial fluctuation range;

[0128] Step S505: If the second deviation degree is less than the second deviation threshold and the drift trend is consistent in direction, then the parameter drift judgment result is determined to be a global drift.

[0129] In some embodiments, global drift may exist due to workpiece material properties, initial clamping, or environmental factors. This drift is not directly caused by tool wear or machine tool component aging. Failure to effectively identify this drift may affect the accuracy of subsequent health baseline generation and local tool damage assessment. Therefore, a multi-cycle cutting test can be performed on the milling disc cutter before workpiece machining to obtain initial impact characteristic data. The multi-cycle cutting test is typically conducted in a controlled environment to ensure that the acquired initial impact characteristic data accurately reflects the cutting characteristics of the tool under undamaged conditions. The initial impact characteristic data may include parameters such as impact force, vibration amplitude, and energy of each tool at the moment of cutting.

[0130] Then, based on the initial impact characteristic data, the initial fluctuation range is determined. The initial fluctuation range characterizes the normal fluctuation range of the instantaneous impact characteristics of the cutting tool under healthy conditions, taking into account factors such as cutting tool manufacturing tolerances, assembly errors, and material micro-inhomogeneities. The initial fluctuation range can be defined by the statistical properties of the initial impact characteristic data (e.g., the mean plus or minus a certain multiple of the standard deviation).

[0131] Next, based on the initial impact characteristic data, the drift trend is calculated. The drift trend refers to the direction and magnitude of change in the initial impact characteristic data over time or the number of cuts during multi-cycle cutting tests. For example, methods such as linear regression or moving average can be used to analyze whether there is a continuous, consistent, small change in the initial impact characteristic data. Then, based on the characteristic change rate and the initial fluctuation range, a second degree of deviation is calculated. The second degree of deviation quantifies the degree of deviation between the current instantaneous impact characteristic and the initial fluctuation range, and incorporates information about the characteristic change rate.

[0132] If the second deviation is less than the second deviation threshold, and the drift trend is consistent in direction, then the parameter drift is determined to be a global drift. The second deviation threshold is a preset boundary used to distinguish between normal fluctuations and abnormal changes, and can be set through expert experience. When the deviation is small, but all or most cutting tools exhibit a consistent drift direction, this usually indicates a global, non-localized parameter change caused by damage, such as batch differences in workpiece material, changes in cutting fluid performance, or minor changes in the overall stiffness of the machine tool.

[0133] To illustrate this technical solution more clearly, a specific example is used below. Suppose that a multi-cycle cutting test is performed on a milling disc cutter before machining a new batch of workpieces. By collecting instantaneous impact characteristic data for each insert, it is found that the impact characteristic values ​​of all inserts are generally slightly higher than the initial baseline of previous batches of workpieces. However, this increase is uniform and consistent in direction, and its deviation is less than a preset second deviation threshold. Simultaneously, analysis of the initial impact characteristic data reveals that its drift trend also exhibits consistent direction. In this situation, the system determines that the current parameter change is a global drift. This may mean that the material hardness of the new batch of workpieces has slightly increased, or that the overall stiffness of the machine tool has slightly adjusted due to changes in ambient temperature, rather than that the insert has suffered localized damage. By identifying this global drift, the system can adjust the health baseline accordingly, avoiding false reports of insert damage and thus ensuring the accuracy of subsequent localized insert damage assessments.

[0134] Through the above technical solution, this embodiment can effectively identify and distinguish global drift, avoiding misjudging parameter changes caused by non-destructive factors as local tool damage, thereby improving the accuracy of tool damage assessment. Simultaneously, by identifying global drift, a more refined basis can be provided for the adaptive adjustment of the health baseline, enabling the health baseline to better adapt to changes in the actual machining environment, further enhancing the robustness and adaptability of the intelligent control system for high-speed milling disc cutters.

[0135] In some embodiments, step S502, determining the initial fluctuation range based on the initial impact characteristic data, may include, but is not limited to, the following steps:

[0136] Step S601: Obtain information on macroscopic defects and microstructural inhomogeneities of the material in the current cutting area;

[0137] Step S602: Determine the upper limit and lower limit of the fluctuation range based on the macroscopic defect information and microstructural inhomogeneity information of the material.

[0138] Step S603: Determine the initial fluctuation range based on the initial impact characteristic data, the upper limit of the fluctuation range, and the lower limit of the fluctuation range.

[0139] In some embodiments, information on macroscopic defects and microstructural inhomogeneities of the material in the current cutting region can be acquired first. Macroscopic defects refer to large-scale inhomogeneities in the workpiece material, such as inclusions, pores, and cracks. These defects can lead to abnormal fluctuations in instantaneous impact characteristics during cutting. Microstructural inhomogeneities refer to differences at the microscopic level, such as grain size, grain boundary distribution, and phase composition. These differences also affect cutting forces and vibration response. This information can be acquired in various ways, such as performing non-destructive testing (e.g., ultrasonic testing, X-ray testing) on ​​the workpiece before machining to identify macroscopic defects, or obtaining microstructural data through metallographic analysis, scanning electron microscopy, etc. The aim is to comprehensively understand the intrinsic properties of the workpiece material and provide fundamental data for determining the subsequent fluctuation range.

[0140] Then, based on information about macroscopic defects and microstructural inhomogeneities in the material, the upper and lower limits of the fluctuation range are determined. These limits represent the normal fluctuation range of instantaneous impact characteristics that may be caused by the material's own inhomogeneities under ideal healthy conditions. For example, when there are inclusions of known size and location in the material, the instantaneous increase or decrease in instantaneous impact characteristics that may occur when the cutting tool passes through that area can be predicted, thereby setting the corresponding fluctuation range.

[0141] Then, based on the initial impact characteristic data, the upper limit of the fluctuation range, and the lower limit of the fluctuation range, the initial fluctuation range is determined. This initial fluctuation range comprehensively considers the inherent impact characteristics of the tool in a healthy state, as well as the normal fluctuations caused by the characteristics of the workpiece material itself, thereby establishing a more accurate and robust benchmark.

[0142] Through the above technical solution, this embodiment can fully consider the inherent characteristics of the workpiece material, including its macroscopic defects and microstructural inhomogeneities. This makes the determined initial fluctuation range closer to the actual cutting conditions, avoiding misjudgments caused by the material's own inhomogeneities. Therefore, subsequent assessment of local tool damage can more accurately distinguish between abnormal fluctuations caused by tool wear or damage and normal fluctuations caused by workpiece material characteristics, significantly improving the accuracy and reliability of tool damage assessment, thus providing a more solid data foundation for the intelligent control of high-speed milling disc cutters.

[0143] In some embodiments, in step S602, determining the upper limit and lower limit of the fluctuation range based on the material's macroscopic defect information and microstructural inhomogeneity information may include, but is not limited to, the following steps:

[0144] Constructing a three-dimensional map of the workpiece's material properties;

[0145] Based on the real-time position and cutting depth of the milling disc cutter, the material property information of the current cutting area is retrieved from the three-dimensional material property map of the workpiece;

[0146] Based on material property information and cutting direction of the cutting tool, calculate the degree of influence of material properties on instantaneous impact characteristics;

[0147] Based on the degree of impact, information on macroscopic defects in the material, and information on microstructural inhomogeneity, the upper and lower limits of the fluctuation range are determined.

[0148] In some embodiments, the macroscopic defects and microstructural inhomogeneities of the workpiece material may differ spatially, and their influence on instantaneous impact characteristics is not constant but closely related to the real-time cutting state (such as position, depth, and direction) of the milling cutter. If the upper and lower limits of the fluctuation range are determined solely based on static or locally acquired material information, it may not adequately reflect the dynamic changes in material properties during cutting and their actual impact on impact characteristics. This results in an inaccurate determination of the initial fluctuation range, affecting the accuracy of subsequent tool damage assessment.

[0149] To this end, a three-dimensional material property map of the workpiece can be constructed first. Before or during processing, a digital model reflecting the distribution of the workpiece's internal material properties (such as hardness distribution, density, grain size, inclusions, porosity, and other macroscopic defects and microstructural inhomogeneities) in three-dimensional space can be established using non-destructive testing techniques (e.g., ultrasonic testing, X-ray diffraction, eddy current testing, etc.) or in combination with a material database and finite element simulation. This map can be regarded as a "digital twin" of the workpiece material, providing basic data for subsequent accurate analysis.

[0150] Then, based on the real-time position and depth of cut of the milling cutter, the system searches for material property information of the current cutting area in the workpiece material property 3D atlas. The machine tool's CNC system can be used to obtain the precise 3D position of the milling cutter in the workpiece coordinate system and the current depth of cut. Based on this data, the system queries the pre-constructed workpiece material property 3D atlas to obtain detailed property information of the material area actually cut by the current insert. For example, it can query the local hardness of the area, the presence of microcracks or inclusions, etc. The purpose is to obtain the local material properties most relevant to the current cutting behavior, rather than general overall material properties.

[0151] Then, based on material property information and the cutting direction of the insert, the degree of influence of material properties on instantaneous impact characteristics is calculated. The material properties of the current cutting area (such as local hardness, toughness, defect type and distribution) and the cutting direction of the insert relative to the material (such as climb milling, conventional milling, entry angle, exit angle, etc.) can be analyzed to assess how these factors collectively affect the instantaneous impact force during the insert cutting process. For example, when the insert cuts into a high-hardness region or encounters inclusions, the instantaneous impact force will increase significantly; different cutting directions will also lead to differences in impact loads. This calculation can be performed using a pre-defined physical model, empirical formula, or machine learning model, with the aim of quantifying the dynamic influence of material properties and cutting direction on instantaneous impact characteristics.

[0152] Finally, based on the degree of impact, information on macroscopic defects in the material, and information on microstructural inhomogeneities, the upper and lower limits of the fluctuation range are determined. A baseline fluctuation range can be established, and dynamically adjusted according to the degree of impact. That is, when the impact is significant, the upper and lower limits of the fluctuation range are appropriately widened to accommodate normal impact fluctuations caused by changes in local material properties; conversely, they are narrowed. The aim is to ensure that the determined upper and lower limits of the fluctuation range more accurately reflect the reasonable fluctuation range of instantaneous impact characteristics caused by the material's inherent properties under current cutting conditions.

[0153] To illustrate this technical solution more clearly, a specific example is used below. Suppose a high-speed milling operation is performed on an aero-engine blade made of a special alloy, which may contain minute grain inhomogeneities, localized hardness differences, or small amounts of non-metallic inclusions. Before machining, the blade is first comprehensively scanned using high-precision non-destructive testing equipment (e.g., ultrasonic phased array or X-ray CT scan), and combined with material mechanical property test data, a detailed three-dimensional material property map of the workpiece is constructed. This map accurately records information such as hardness values, grain structure, defect locations, and sizes at various points inside the blade. During the actual milling process, the machine tool's CNC system monitors the tip position and depth of cut of the milling cutter in real time. For example, when a cutting edge of the milling cutter cuts into a specific area inside the blade, the system immediately queries the three-dimensional map for the material property information of that area, such as finding that the local hardness of that area is 10% higher than the average and that a small carbide inclusion is present. At the same time, the system will also obtain the cutting direction of the current cutting blade cutting this area (for example, whether it is cutting with or against the grain direction).

[0154] Next, based on a pre-established physical model and empirical data, the system calculates the potential impact of such localized high-hardness regions and carbide inclusions on the instantaneous impact characteristics under the current cutting direction. For example, calculations show that due to increased local hardness and the presence of inclusions, the amplitude of the instantaneous impact characteristics may increase by 15% to 20% from the normal range. Finally, this calculated impact level is integrated with the original macroscopic material defect information and microstructural inhomogeneity information to dynamically adjust the upper and lower limits of the initial fluctuation range. For example, if the normal fluctuation range is ±5%, then when encountering the aforementioned high-hardness inclusion region, the fluctuation range may be dynamically adjusted to +25% / -5% to ensure that normal impact fluctuations caused by material properties are included and not misjudged as tool damage. In this way, even on workpieces with complex and variable material properties, the accuracy of the health baseline can be ensured, thereby improving the reliability of tool damage assessment.

[0155] Through the above technical solution, this embodiment can achieve a more accurate and dynamic determination of the initial fluctuation range. By fully considering the spatial distribution of workpiece material properties and the dynamic influence of the cutting process, the determined upper and lower limits of the fluctuation range can more realistically reflect the actual material condition of the current cutting area. This significantly improves the accuracy and robustness of health baseline generation, enabling subsequent local tool damage assessment to more effectively eliminate interference caused by the inhomogeneity of the workpiece material itself. This enhances the sensitivity and reliability of tool damage assessment, reduces the risk of false alarms and missed alarms, and provides a more solid data foundation for the intelligent control of high-speed milling disc cutters.

[0156] The beneficial effects of implementing the embodiments of the present invention include: First, the state signal during the cutting process is acquired, and the state signal is time-synchronized with the cutting action of the milling disc cutter to obtain a synchronization correlation signal. Then, based on the synchronization correlation signal, the instantaneous impact characteristics corresponding to the cutting moment of each insert in the milling disc cutter are extracted, and multiple instantaneous impact characteristics are periodically accumulated and statistically analyzed to generate a health baseline. Then, based on the current vibration signal, instantaneous impact characteristics, and health baseline, the local damage of the insert is judged to obtain the local damage judgment result of the insert. Finally, based on the local damage judgment result of the insert, graded protection control is performed, thereby enabling the judgment of local damage of the insert by combining instantaneous impact characteristics and health baseline, so as to realize intelligent control of the milling disc cutter and improve accuracy and reliability.

[0157] like Figure 2 As shown, this embodiment of the invention also provides an intelligent control system for high-speed milling disc cutters, including:

[0158] The information acquisition module 701 is used to acquire the status signals during the cutting process, including the load signal and the current vibration signal.

[0159] The time synchronization association module 702 is used to synchronize the status signal with the cutting action of the milling disc cutter to obtain a synchronization association signal.

[0160] The feature extraction module 703 is used to extract the instantaneous impact features corresponding to the cutting moment of each insert in the milling disc cutter based on the synchronous correlation signal;

[0161] The baseline generation module 704 is used to periodically accumulate and statistically analyze multiple instantaneous impact characteristics to generate a healthy baseline;

[0162] The blade damage judgment module 705 is used to judge the local damage of the blade based on the current vibration signal, instantaneous impact characteristics and health baseline, and obtain the blade local damage judgment result.

[0163] The protection control module 706 is used to perform graded protection control based on the judgment results of local damage to the blade.

[0164] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0165] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

Claims

1. A method for intelligent control of high-speed milling disc cutters, characterized in that, Includes the following steps: Acquire state signals during the cutting process, including load signals and current vibration signals; The state signal is time-synchronized with the cutting action of the milling disc cutter to obtain a synchronization correlation signal; Based on the synchronous correlation signal, extract the instantaneous impact characteristics corresponding to the cutting moment of each insert in the milling disc cutter; A health baseline is generated by periodically accumulating and statistically analyzing multiple instantaneous impact characteristics. Based on the current vibration signal, the instantaneous impact characteristics, and the health baseline, a local damage assessment of the blade is performed to obtain the local damage assessment result of the blade. Based on the results of the assessment of local damage to the blade, graded protection and control are implemented. The step of determining local blade damage based on the current vibration signal, the instantaneous impact characteristics, and the health baseline, and obtaining the local blade damage determination result, includes: Calculate the first degree of deviation between the instantaneous impact characteristic and the health baseline; If the first deviation is greater than the first deviation threshold, then determine whether there is a periodic correlation between the first deviation and the spindle speed; If the first deviation degree is periodically correlated with the spindle speed, then frequency analysis is performed on the current vibration signal to obtain the frequency analysis results; Based on the frequency analysis results, the target vibration is determined, including conventional machining vibration and periodic low-frequency vibration caused by tool damage; Based on the first degree of deviation and the target vibration, a local damage assessment result for the blade is generated.

2. The method according to claim 1, characterized in that, The process of periodically accumulating and statistically analyzing multiple instantaneous impact characteristics to generate a health baseline includes: Identify the current group of cutting blades within the current cutting area; Based on the instantaneous impact characteristics of each current blade in the current blade group, a first statistical feature is calculated, the first statistical feature including a first mean and a first standard deviation; Identify the minimum fluctuation range based on the instantaneous impact characteristics of each current blade in the current blade group; Determine the range of health fluctuations based on the minimum fluctuation range; If the first statistical feature satisfies the health fluctuation range, then the health baseline is generated based on the first statistical feature and the health fluctuation range. If the first statistical feature does not meet the health fluctuation range, then baseline adaptive adjustment processing is performed based on the current blade group to generate the health baseline.

3. The method according to claim 2, characterized in that, The process of generating the healthy baseline by performing baseline adaptive adjustment based on the current blade group includes: Based on the current blade group, identify neighboring blade groups within the immediate cutting area; Based on the instantaneous impact characteristics of each neighboring blade in the neighboring blade group, a second statistical feature is calculated, which includes a second mean and a second standard deviation. The health fluctuation range is updated based on the second statistical feature; The health baseline is generated based on the second statistical feature and the updated health fluctuation range.

4. The method according to claim 3, characterized in that, The step of generating the health baseline based on the second statistical feature and the updated health fluctuation range includes: Calculate the trend of feature change based on the first statistical feature and the second statistical feature; Calculate the rate of change of the features based on the described trend of change; Based on the rate of change of the described features, parameter drift is determined, and the parameter drift determination result is obtained. Based on the parameter drift judgment result, the first statistical feature and the second statistical feature are weighted and fused to obtain the target statistical feature; The health baseline is generated based on the target statistical characteristics and the updated health fluctuation range.

5. The method according to claim 4, characterized in that, The step of determining parameter drift based on the characteristic change rate and obtaining the parameter drift determination result includes: Acquire historical vibration signals of a target component, which includes a machine tool spindle bearing or guide rail; The initial vibration frequency and initial amplitude were obtained by performing spectral analysis on the historical vibration signal. Based on the initial vibration frequency and the initial amplitude, an initial vibration characteristic spectrum is constructed; Monitor the current vibration signal of the target component; Perform spectral analysis on the current vibration signal to obtain the current vibration frequency and current amplitude; Based on the current vibration frequency and the current amplitude, construct the current vibration feature map; The initial vibration feature spectrum and the current vibration feature spectrum are compared to obtain the comparison result; If the comparison result shows that there is no localized and non-unidirectional slow drift, and the feature change rate is less than the preset wear change threshold, then the parameter drift judgment result is determined to be uniform wear of the tool as a whole. If the comparison result shows a slow drift with locality and non-uniformity, and the rate of change of the feature is greater than a preset wear change threshold, then the parameter drift judgment result is determined to be a non-uniform drift caused by the aging of machine tool components.

6. The method according to claim 4, characterized in that, The step of determining parameter drift based on the characteristic change rate and obtaining the parameter drift determination result includes: Before machining the workpiece, a multi-cycle cutting test was performed on the milling disc cutter to obtain initial impact characteristic data; Based on the initial impact characteristic data, determine the initial fluctuation range; Calculate the drift trend based on the initial impact characteristic data; The second degree of deviation is calculated based on the characteristic change rate and the initial fluctuation range; If the second deviation degree is less than the second deviation threshold, and the drift trend is consistent in direction, then the parameter drift judgment result is determined to be a global drift.

7. The method according to claim 6, characterized in that, Determining the initial fluctuation range based on the initial impact characteristic data includes: Obtain information on macroscopic defects and microstructural inhomogeneities of the material in the current cutting area; Based on the macroscopic defect information and the microstructural inhomogeneity information of the material, the upper limit and lower limit of the fluctuation range are determined; The initial fluctuation range is determined based on the initial impact characteristic data, the upper limit of the fluctuation range, and the lower limit of the fluctuation range.

8. The method according to claim 7, characterized in that, The step of determining the upper limit and lower limit of the fluctuation range based on the macroscopic defect information and the microstructural inhomogeneity information of the material includes: Constructing a three-dimensional map of the workpiece's material properties; Based on the real-time position and cutting depth of the milling disc cutter, the material property information of the current cutting area is retrieved from the three-dimensional material property map of the workpiece; Based on the material property information and the cutting direction of the blade, calculate the degree of influence of the material properties on the instantaneous impact characteristics; Based on the degree of influence, the macroscopic defect information of the material, and the microstructural inhomogeneity information, the upper limit of the fluctuation range and the lower limit of the fluctuation range are determined.

9. A high-speed milling disc cutter intelligent control system, characterized in that, include: The information acquisition module is used to acquire status signals during the cutting process, including load signals and current vibration signals; The time synchronization association module is used to synchronize the status signal with the cutting action of the milling disc cutter to obtain a synchronization association signal; The feature extraction module is used to extract the instantaneous impact features corresponding to the cutting moment of each insert in the milling disc cutter based on the synchronous correlation signal; The baseline generation module is used to periodically accumulate and statistically analyze multiple instantaneous impact characteristics to generate a healthy baseline; The blade damage assessment module is used to assess local blade damage based on the current vibration signal, the instantaneous impact characteristics, and the health baseline, and to obtain the blade local damage assessment result. The protection control module is used to perform graded protection control based on the judgment result of local damage to the blade; The step of determining local blade damage based on the current vibration signal, the instantaneous impact characteristics, and the health baseline, and obtaining the local blade damage determination result, includes: Calculate the first degree of deviation between the instantaneous impact characteristic and the health baseline; If the first deviation is greater than the first deviation threshold, then determine whether there is a periodic correlation between the first deviation and the spindle speed; If the first deviation degree is periodically correlated with the spindle speed, then frequency analysis is performed on the current vibration signal to obtain the frequency analysis results; Based on the frequency analysis results, the target vibration is determined, including conventional machining vibration and periodic low-frequency vibration caused by tool damage; Based on the first degree of deviation and the target vibration, a local damage assessment result for the blade is generated.

Images