Numerical control lathe hardware machining process control method and system

Abstract

The embodiment of the invention provides a numerical control lathe hardware machining process control method and system, and relates to the technical field of numerical control lathe machining process control. Based on a preset judgment rule, performing preliminary abnormality judgment on the sensor signal, and generating a preliminary alarm; receiving feedback information of a user on the preliminary alarm and acquiring sensor signal characteristics and processing parameters related to preliminary abnormality judgment; and according to the feedback information, the sensor signal characteristics and the processing parameters, performing abnormity judgment and alarm generation. The overall utilization rate and the machining efficiency of the equipment can be improved.

Description

Technical Field

[0001] This application relates to the field of CNC lathe machining process control technology, and in particular to a method and system for controlling the machining process of hardware parts on a CNC lathe. Background Technology

[0002] In modern industrial production, the machining accuracy and efficiency of CNC lathe hardware parts are key indicators for measuring production capacity. To ensure the stability of the machining process and product quality, many advanced manufacturing enterprises have introduced real-time monitoring technology. This real-time monitoring method for controlling the machining process of CNC lathe hardware parts aims to detect potential anomalies in a timely manner by continuously collecting various data during the machining process, such as cutting force, vibration, and temperature, and adjust machining parameters or issue early warnings accordingly, thereby avoiding scrap, extending tool life, and improving overall production efficiency. However, in actual industrial production environments, equipment maintenance and parts replacement are unavoidable daily operations. For example, when the force sensor mounting bracket of a CNC lathe needs to be replaced due to long-term use, if a non-original manufacturer's replacement bracket is selected, even if it performs normally in static testing, under high-intensity, dynamic machining conditions, subtle differences in its material composition, inherent rigidity, damping characteristics, or manufacturing process compared to the original part may lead to unexpected periodic structural resonance under specific machining parameters (such as spindle speed and depth of cut). This resonance is transmitted to the force sensor through mechanical coupling, causing additional noise generated by the resonance of the support itself to be mixed into the signal collected by the sensor. Summary of the Invention

[0003] This application aims to address at least one of the technical problems existing in the prior art. To this end, this application proposes a method and system for controlling the machining process of hardware parts on a CNC lathe, aiming to improve the overall uptime and machining efficiency of the equipment.

[0004] In a first aspect, embodiments of this application provide a method for controlling the machining process of hardware parts on a CNC lathe, including: Acquire sensor signals during the machining process of hardware parts on a CNC lathe; Based on preset judgment rules, the sensor signals are initially judged for abnormalities and an initial alarm is generated; Receive user feedback on initial alarms and acquire sensor signal characteristics and processing parameters related to initial anomaly assessment; Based on feedback information, sensor signal characteristics, and processing parameters, anomaly detection and alarm generation are performed.

[0005] According to some embodiments of this application, anomaly detection and alarm generation are performed based on feedback information, sensor signal characteristics, and processing parameters, including: Obtain background vibration and environmental information of the machine tool related to the initial anomaly assessment; Based on feedback information, sensor signal characteristics, processing parameters, machine tool background vibration, and environmental information, a context-related information package is obtained. When a context-related information packet appears, the similarity between the context-related information packet and the context-related information packets corresponding to different intent tags in the preset learning sample library is calculated. The preset learning sample library is established based on historical anomaly types and historical context-related information packets. Calculate the confidence level of the current anomaly attribution based on similarity; Anomaly detection and alert generation are performed based on the confidence level of the current anomaly attribution.

[0006] According to some embodiments of this application, when a context-related information packet appears, the similarity between the context-related information packet and context-related information packets corresponding to different intent tags in a preset learning sample library is calculated, including: Identify features in context-related information packets that are associated with frequency drift or dynamic evolution; By comparing the dynamic evolution features of context-related information packages with the dynamic evolution features of context-related information packages corresponding to different intent labels in the preset learning sample library, the matching results of dynamic evolution features are obtained. Based on the matching results of the dynamic evolution features, the similarity between the new context-related information package and the context-related information packages corresponding to different intent labels in the learning sample library is calculated.

[0007] According to some embodiments of this application, calculating the confidence level of the current anomaly attribution based on similarity includes: Compare the differences between the similarity between the context-related information package and the tool wear pattern, and the similarity between the context-related information package and the interface micro-degradation pattern, and the preset discrimination thresholds. When the difference is less than the preset discrimination threshold, the occurrence frequency of high-frequency transient pulses and the rate of curvature change of low-frequency baseline drift in the analysis context association information packet are analyzed. The confidence level of the current anomaly attribution is calculated based on the comparison between the occurrence frequency of the high-frequency transient pulse and the preset pulse frequency threshold, and the comparison between the curvature change rate of the low-frequency baseline drift and the preset curvature threshold.

[0008] According to some embodiments of this application, anomaly judgment and alarm generation are performed based on the confidence level of the current anomaly attribution, including: Based on the confidence level of the current anomaly attribution, multiple candidate avoidance suggestions are selected from the preset avoidance strategy library; For each candidate avoidance suggestion, the impact of the candidate avoidance suggestion on short-term production efficiency, tool life, equipment health and overall production plan is simulated to obtain multiple simulation impact result scores; The scores of multiple simulation impact results are weighted and evaluated to calculate the comprehensive benefit score; The avoidance suggestion with the highest overall benefit score is selected as the final anomaly assessment and alarm generation output.

[0009] According to some embodiments of this application, a comprehensive benefit score is calculated by weighting and evaluating multiple simulation impact scores, including: Obtain the urgency of the current machining task, the historical equipment failure rate, and the actual wear of the cutting tools; Based on the degree of urgency, the weight of short-term production efficiency in the calculation of the comprehensive benefit score is adjusted to obtain the first weight; Based on the historical failure rate of the equipment, the weight of the equipment health status in the calculation of the comprehensive benefit score is adjusted to obtain the second weight; Based on the actual wear of the tool, the weight of tool life in the calculation of the comprehensive benefit score is adjusted to obtain the third weight; Based on the level of urgency, the weight of production batch delivery time in the calculation of the overall benefit score is adjusted to obtain the fourth weight; The comprehensive benefit score is calculated by weighting the scores of multiple simulated impact results according to the first weight, the second weight, the third weight, and the fourth weight.

[0010] According to some embodiments of this application, the weight of short-term production efficiency in the calculation of the comprehensive benefit score is adjusted according to the degree of urgency to obtain a first weight, including: Acquire historical production efficiency fluctuations, tool wear rates, and equipment micro-vibration patterns; Analyze the correlation between the level of urgency and historical production efficiency fluctuations, tool wear rate, and equipment micro-vibration modes; Based on the correlation, the potential impact of improved production efficiency on tool life is predicted, and the prediction results are obtained. Based on the forecast results, the weight of short-term production efficiency in the calculation of the comprehensive benefit score is adjusted to obtain the first weight.

[0011] According to some embodiments of this application, the correlation between urgency level and historical production efficiency fluctuations, accelerated tool wear, and equipment micro-vibration modes is analyzed, including: Obtain information on the material type and geometry of the hardware parts currently being processed; Based on the material type and geometric shape information of the hardware parts, select the corresponding material-geometry specific association mode from the preset association mode library; Based on the material-geometry specific correlation model, the correlation between urgency and historical production efficiency fluctuations, tool wear acceleration, and equipment micro-vibration modes is analyzed.

[0012] According to some embodiments of this application, based on correlation, the potential impact of increased production efficiency on tool life is predicted, and the prediction results are obtained, including: Obtain the type of tool material, tool coating type, and cutting fluid used in the current machining process; Based on the tool material type, tool coating type, and cutting fluid type, select the corresponding wear mechanism model from the preset tool wear mechanism library; The interaction between the urgency level and various wear factors in the wear mechanism model was analyzed, and the results of the interaction analysis were obtained. Based on the interaction analysis results, the influence weights of each wear factor on the tool wear rate in the wear mechanism model were adjusted to obtain the adjusted wear mechanism model. Based on the adjusted wear mechanism model and correlation, the potential impact of improved production efficiency on tool life is predicted, and the prediction results are obtained.

[0013] Secondly, embodiments of this application provide a CNC lathe hardware machining process control system, including: The acquisition module is used to acquire sensor signals during the machining process of hardware parts on a CNC lathe; The generation module is used to perform preliminary anomaly judgment on sensor signals based on preset judgment rules and generate preliminary alarms; The data logging module is used to receive user feedback on the initial alarm and to acquire sensor signal characteristics and processing parameters related to the initial anomaly judgment. The judgment and generation module is used to judge anomalies and generate alarms based on feedback information, sensor signal characteristics, and processing parameters.

[0014] According to the technical solution of this application, at least the following beneficial effects are achieved: By introducing user feedback information and combining it with sensor signal characteristics and machining parameters for comprehensive judgment, this application can perform secondary verification and correction of preliminary alarms. User feedback information provides supplementary information based on human experience and actual conditions, helping the system distinguish between interference noise caused by hardware mismatch and genuine tool wear signals. For example, when the system initially judges it as tool wear, but the user reports that the actual machining effect is good or the tool has been replaced recently, the system can combine this information to reassess the authenticity of the anomaly. At the same time, acquiring sensor signal characteristics and machining parameters related to the preliminary anomaly judgment allows the system to analyze the specific context in which the anomaly occurs more deeply, such as whether a certain type of anomaly signal is more likely to occur at a specific spindle speed and depth of cut. Through this multi-dimensional, contextualized judgment mechanism, this application can significantly improve the accuracy of anomaly judgment and effectively avoid frequent and unnecessary tool replacements caused by misjudgment. This can not only significantly reduce tool costs and inventory pressure, reduce production line interruptions, and improve the overall uptime and machining efficiency of equipment, but also overcome the negative production impact caused by the fragility of decision logic in the real-time monitoring system of the prior art, achieving more stable and efficient CNC lathe hardware machining process control.

[0015] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0016] The accompanying drawings are used to provide a further understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0017] Figure 1 A flowchart illustrating a method for controlling the machining process of hardware parts on a CNC lathe, as provided in one embodiment of this application; Figure 2 This is a schematic diagram of a CNC lathe hardware machining process control system provided in one embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical methods, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0019] It should be noted that the meaning of "multiple" (or "more than") in the description of the embodiments of this application refers to two or more, and "greater than," "less than," "exceeding," etc. are understood to exclude the number itself, while "above," "below," "within," etc. are understood to include the number itself. If "first," "second," etc. are used in the description, they are only for the purpose of distinguishing technical features and should not be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated or the order of the technical features indicated.

[0020] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, the simultaneous existence of A and B, or the existence of B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, and c can represent: the existence of a alone, the existence of b alone, the existence of c alone, the simultaneous existence of a and b, the simultaneous existence of a and c, the simultaneous existence of b and c, or the simultaneous existence of a, b, and c, where a, b, and c can be single or multiple.

[0021] In the description of this application, unless otherwise expressly defined, terms such as "setup," "installation," and "connection" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this application in conjunction with the specific content of the technical solution.

[0022] Based on the above, this application proposes a method and system for controlling the machining process of hardware parts on a CNC lathe, aiming to improve the overall uptime and machining efficiency of the equipment.

[0023] The CNC lathe hardware parts machining process control method provided in this application embodiment can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, and can also be configured to provide cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, and content delivery networks (CDNs). Cloud servers providing basic cloud computing services such as big data and artificial intelligence platforms; software can be applications that implement process control methods for CNC lathe hardware parts processing, but is not limited to the above forms.

[0024] This application can be applied to numerous general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics devices, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via communication networks. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices. It should be noted that in various specific embodiments of this invention, when processing is required based on data related to the characteristics of an object (e.g., user attributes or sets of attribute information), permission or consent from the corresponding object is obtained first, and the collection, use, and processing of this data comply with relevant laws and standards. Furthermore, when the embodiments of the present invention need to obtain the attribute information of an object, they will obtain the separate permission or separate consent of the corresponding object through pop-up windows or redirection to a confirmation page. After obtaining the separate permission or separate consent of the corresponding object, they will then obtain the relevant data of the object necessary for the embodiments of the present invention to operate normally.

[0025] See Figure 1 , Figure 1 This is a flowchart illustrating a method for controlling the machining process of hardware parts on a CNC lathe, as provided in one embodiment of this application. The method for controlling the machining process of hardware parts on a CNC lathe, as provided in this embodiment, includes, but is not limited to, the following steps... Up to step S140 The following sections will describe each step in turn.

[0026] step Acquire sensor signals during the machining process of hardware parts on a CNC lathe; step Based on preset judgment rules, perform preliminary anomaly judgment on sensor signals and generate preliminary alarms; step It receives user feedback on initial alarms and acquires sensor signal characteristics and processing parameters related to initial anomaly assessment. step Based on feedback information, sensor signal characteristics, and processing parameters, anomaly detection and alarm generation are performed.

[0027] It should be noted that "sensor signals" refer to the physical quantity data collected in real time by various sensors (such as vibration sensors, acoustic emission sensors, force sensors, temperature sensors, etc.) during the machining of metal parts on CNC lathes. These signals are the basic data reflecting the machining status and equipment health. "Preset judgment rules" refer to a series of logical conditions or thresholds pre-set before the system is put into use, based on a large amount of historical machining data, expert experience, and specifications provided by the equipment manufacturer, to identify machining anomalies. For example, when the root mean square value of the vibration signal exceeds a certain threshold, it is initially judged as an anomaly. "Preliminary alarm" refers to the first-level anomaly prompt issued by the system after preliminary analysis of the sensor signals according to the preset judgment rules. This alarm is usually automatically generated and aims to quickly attract the operator's attention. "Feedback information" refers to the confirmation, correction, or supplementary explanation provided by the user (usually an operator or engineer) after receiving the preliminary alarm, based on their on-site experience, observations, or further testing, regarding the authenticity, anomaly type, or severity of the preliminary alarm. For example, the operator might provide feedback such as "the alarm is a false alarm, there is actually no anomaly" or "the alarm is true, the anomaly type is tool wear." "Sensor signal characteristics" refer to key information extracted from raw sensor signals that characterizes the machining state or anomaly type, such as the signal's frequency components, amplitude, energy, and time-domain statistics (e.g., mean, variance, kurtosis, skewness). "Machining parameters" refer to various process parameters set during CNC lathe machining, such as spindle speed, feed rate, depth of cut, width of cut, tool type, and workpiece material. These parameters directly affect the physical phenomena of the machining process and the characteristics of the sensor signals. "Anomaly judgment" refers to the system's deeper and more accurate assessment and identification of whether anomalies exist, their type, and their severity during machining, after comprehensively considering multi-source information such as sensor signals, machining parameters, and user feedback. "Alarm generation" refers to the system issuing alarm information with higher confidence and more specific guidance to the user after anomaly judgment, which may include the anomaly type and suggested avoidance measures.

[0028] In one embodiment, in the step of "acquiring sensor signals during the machining process of CNC lathe hardware parts," the system continuously collects data from various sensors installed on the CNC lathe. For example, piezoelectric accelerometers can be used to monitor the vibration of the machine tool, acoustic emission sensors can be used to capture high-frequency transient signals during the cutting process, or thermocouples can be used to measure the temperature of the cutting area. These sensor signals can be analog signals, converted into digital signals by an analog-to-digital converter (ADC) before being transmitted to the data processing unit; or they can be intelligent sensors that directly output digital signals. For example, when machining a stainless steel hardware part, the system will acquire the vibration data of the spindle bearing, the acoustic emission signal of the tool-workpiece contact area, and the temperature data of the cutting fluid in real time. This raw data forms the basis for all subsequent judgments. Secondly, in the step of "based on preset judgment rules, performing preliminary anomaly judgment on sensor signals and generating preliminary alarms," ​​the system will perform preliminary and rapid analysis on the acquired sensor signals. For example, a vibration amplitude threshold can be set. When the signal amplitude of any vibration sensor continuously exceeds the threshold for a certain period of time (e.g., 50 milliseconds), the system determines it as a preliminary anomaly. Alternatively, an acoustic emission signal energy threshold can be set. When the average energy of the acoustic emission signal suddenly rises above a preset value within a certain time window, a preliminary anomaly is also triggered. These preset judgment rules are usually based on experience or simple statistical models, aiming to quickly capture potential anomalies. When these rules are met, the system will immediately generate a preliminary alarm, such as displaying "Vibration abnormality, please pay attention!" on the operation interface or alerting the operator through an audio-visual signal. Next, in the stage of "receiving user feedback on the preliminary alarm and acquiring sensor signal characteristics and processing parameters related to the preliminary anomaly judgment," the system will wait for and receive feedback from the operator on the preliminary alarm. For example, after the system issues a preliminary alarm, the operator may input feedback information such as "false alarm, no actual anomaly," "tool wear," or "workpiece clamping loose" through the human-machine interface based on on-site observation. At the same time, the system will further extract more detailed sensor signal characteristics related to the moment of the preliminary anomaly judgment. For example, if the preliminary alarm is triggered based on vibration signals, the system will extract the spectral characteristics and time-domain statistical characteristics (such as peak factor and kurtosis) of the vibration signals within that time period. In addition, the system records all relevant machining parameters at the time of the initial anomaly, including spindle speed, feed rate, depth of cut, tool type, and workpiece material batch. This multi-dimensional data provides richer information for subsequent accurate judgment. Finally, in the stage of "anomaly judgment and alarm generation based on feedback information, sensor signal characteristics, and machining parameters," the system comprehensively analyzes the previously collected user feedback information, detailed sensor signal characteristics, and machining parameters.For example, if the initial alarm is about abnormal vibration, and the user reports "tool wear," the system will combine the current vibration signal spectrum characteristics (such as an increase in specific wear frequency components) and machining parameters (such as increased cutting force) to verify the accuracy of the user's feedback. If the user reports a "false alarm," the system will analyze whether the current sensor signal characteristics and machining parameters do not actually conform to any known abnormal patterns. Through this fusion and cross-validation of multi-source information, the system can more accurately determine the true type and cause of the anomaly, thereby avoiding false alarms or missed alarms. Based on the final anomaly judgment result, the system will generate a more accurate and instructive alarm, such as "Confirmed as slight tool wear; it is recommended to replace the tool after the next workpiece is machined," or "Confirmed as equipment resonance; it is recommended to check the fixture or adjust the machining parameters."

[0029] It should be noted that acquiring machine tool background vibration and environmental information related to the initial anomaly assessment refers to the system actively or passively collecting background vibration data and workshop environmental data during or after the initial anomaly assessment. Machine tool background vibration can be understood as the inherent vibration generated by the machine tool's own structure, transmission system, cooling system, etc., under no-cutting or stable cutting conditions. This data can be acquired through vibration sensors installed on the machine tool body. Environmental information may include workshop temperature, humidity, air pressure, etc., which can be monitored in real time through environmental sensors. Acquiring this information aims to provide a more comprehensive context for subsequent anomaly assessments. Based on feedback information, sensor signal characteristics, machining parameters, machine tool background vibration, and environmental information, a context-related information package is obtained. This context-related information package is a comprehensive data structure that integrates user feedback on the initial alarm, sensor signal characteristics related to the initial anomaly assessment (e.g., spectral characteristics of vibration signals, intensity changes of acoustic signals), current machining parameters (e.g., spindle speed, feed rate, depth of cut), and newly acquired machine tool background vibration and environmental information. When a context-related information packet appears, the similarity between this packet and context-related information packets corresponding to different intent tags in a pre-set learning sample library is calculated. The pre-set learning sample library is a trained and accumulated knowledge base that stores a large number of historical anomaly types and their corresponding context-related information packets. Each historical anomaly type is assigned an "intent tag," such as "tool wear," "spindle failure," or "loose workpiece clamping." The similarity calculation aims to quantify the degree of matching between the current context-related information packet and known anomaly patterns in the library through machine learning or pattern recognition algorithms, thereby initially identifying the potential attribution of the current anomaly. The confidence level of the current anomaly attribution is then calculated based on the similarity. Confidence level is an indicator of the reliability of the current anomaly attribution, and its calculation can be based on factors such as the magnitude of the similarity, the characteristics of the similarity distribution, and the false positive rate in historical data. For example, when the current context-related information packet has a very high similarity to a sample packet corresponding to a certain intent tag, the confidence level of that intent tag as an anomaly attribution will be high. Therefore, anomaly judgment and alarm generation are performed based on the confidence level of the current anomaly attribution. This means that the final anomaly determination is no longer based solely on preliminary rule-based judgments or user feedback, but rather incorporates deeper contextual analysis and confidence assessment. When the confidence level of an anomaly attribution reaches a preset threshold, the system will confirm the anomaly and generate a corresponding alert. The alert content may include the anomaly type, possible causes, and suggested countermeasures.

[0030] In one embodiment, assuming that during CNC lathe machining, a vibration sensor detects an abnormal high-frequency vibration signal, a preliminary judgment is made that an anomaly may exist. First, the system acquires background vibration data of the machine tool and workshop environmental information related to this preliminary anomaly judgment. For example, it records the vibration spectrum of the machine tool under no-load conditions, as well as the real-time temperature and humidity of the workshop.

[0031] Next, the system integrates user feedback on the initial alarm (e.g., user confirmation of "feeling something abnormal"), characteristics of high-frequency vibration signals (e.g., a significant increase in energy at a specific frequency), current machining parameters (e.g., cutting speed, feed rate), and newly acquired machine tool background vibration and environmental information to form a context-related information package. This context-related information package is then input into a preset learning sample library for comparison. This sample library may contain various historical anomaly patterns such as "tool wear," "spindle bearing failure," and "workpiece clamping looseness," along with their corresponding context-related information packages. The system calculates the similarity between the current context-related information package and these historical patterns. For example, if the current information package is highly similar to the "tool wear" pattern in terms of vibration spectrum characteristics, machining parameter change trends, and user feedback, but less similar to the "spindle bearing failure" pattern, the system calculates the confidence level that the current anomaly is attributed to "tool wear." If this confidence level reaches a preset threshold (e.g., 90%), the system ultimately determines it to be "tool wear" and generates a corresponding alarm, recommending that the operator check or replace the tool.

[0032] It should be noted that identifying features related to frequency drift or dynamic evolution in context-related information packets refers to extracting time-varying frequency components or pattern evolution trends from sensor signal features using signal processing techniques such as wavelet transform, Fourier transform, or time-frequency analysis. The aim is to capture non-static, dynamically changing features in abnormal signals. These features are often key indicators for distinguishing different anomaly types, especially complex faults such as early wear, loosening, or resonance. Specifically, comparing the dynamic evolution features of the context-related information packet with the dynamic evolution features of context-related information packets corresponding to different intent labels in a pre-set learning sample library to obtain the matching results of dynamic evolution features can be understood as comparing the dynamic feature sequence extracted from the current context-related information packet with dynamic feature templates stored in the pre-set learning sample library that are associated with different historical anomaly types (i.e., intent labels). For example, dynamic time warping can be used. Algorithms, Hidden Markov Models (HMMs) or Recurrent Neural Networks Methods such as [list of methods] are used to assess the similarity and consistency of evolution paths between the current dynamic evolution pattern and known anomalous patterns over time. The aim is to accurately quantify the degree of agreement between the dynamic behavior of the current anomalous pattern and known anomalous patterns. In practical applications, calculating the similarity between the new context-related information package and context-related information packages corresponding to different intent labels in the learning sample library, based on the matching results of dynamic evolution features, involves fusing the similarity calculation results of dynamic evolution features with those of other static features (such as amplitude, mean, variance, etc.). For example, ensemble learning methods such as weighted averaging, decision trees, or support vector machines can be used to comprehensively evaluate similarity indicators across different dimensions, thereby obtaining a more comprehensive and accurate overall similarity score. The goal is to ensure that similarity calculations not only consider static features but also effectively reflect the dynamic evolution process of the anomalous pattern, improving the accuracy of anomalous attribution.

[0033] In one embodiment, it is assumed that sensor signals exhibit abnormalities during CNC lathe machining. First, the system extracts features related to frequency drift or dynamic evolution from the current context-related information package. For example, by performing continuous wavelet transform on the vibration signal, its time-frequency plot can be obtained, and the drift trajectory of the dominant frequency component or the energy change trend of specific harmonics can be identified. Subsequently, these extracted dynamic evolution features, such as frequency drift curves or energy spectral density evolution sequences, are compared with historical dynamic evolution patterns stored in a preset learning sample library, which are associated with different intent labels such as "tool wear" and "spindle bearing looseness". For example, if the current frequency drift curve highly matches the historical curve of the "tool wear" pattern in shape and trend, but has a low degree of match with the "spindle bearing looseness" pattern, the matching result of the dynamic evolution features will tend to be "tool wear". Finally, the matching results of these dynamic evolution features are comprehensively evaluated with the similarity calculation results of other static features (such as vibration amplitude, average temperature, etc.) in the context-related information package. For example, this can be fused using a pre-trained classifier to obtain the final similarity score attributing the current anomaly to "tool wear". This method ensures accurate differentiation based on differences in dynamic evolution features, even when static features are similar, thereby improving the accuracy of anomaly detection.

[0034] It should be noted that when calculating the confidence level of the current anomaly attribution, it is first necessary to compare the differences between the similarity between the context-related information package and the tool wear pattern, and the similarity between the context-related information package and the interface micro-degradation pattern, and the preset discrimination threshold. Tool wear pattern and interface micro-degradation pattern refer to two common anomaly types that may exhibit similar sensor signal characteristics during CNC lathe machining of metal parts. Tool wear pattern typically manifests as gradual or abrupt changes in signals such as cutting force, vibration, and acoustic emission, while interface micro-degradation pattern may involve minor loosening, friction, or wear at the spindle-tool interface, workpiece-fixture interface, etc., and its signal characteristics may overlap with tool wear. The preset discrimination threshold is a pre-set value used to determine whether the similarity difference between the two anomaly patterns is large enough for direct differentiation. For example, when the similarity difference between the two patterns is less than this threshold, it indicates that it is difficult to clearly distinguish them based on overall similarity alone, and deeper analysis is required. When the difference is less than the preset discrimination threshold, it indicates that the current anomaly may simultaneously exhibit characteristics of tool wear and interface micro-degradation, or that the macroscopic manifestations of the two modes are too similar. In this case, further analysis of more subtle features in the context-related information package is needed. Specifically, the frequency of high-frequency transient pulses and the rate of change of curvature of low-frequency baseline drift will be analyzed in the context-related information package. High-frequency transient pulses typically refer to short-duration, high-energy spike signals appearing in sensor signals, such as impact vibrations or acoustic emission signals caused by tool micro-chipping, chip impact, or interface fretting friction. Their frequency can reflect the density of the anomaly. Low-frequency baseline drift refers to the slow, continuous change of the sensor signal baseline (average value or trend line), such as the overall increase in cutting force or vibration level caused by gradual tool wear, or the change in system stiffness caused by gradual interface loosening. Its rate of change of curvature can reflect the acceleration or deceleration trend of this drift. Subsequently, based on the comparison between the occurrence frequency of high-frequency transient pulses and a preset pulse frequency threshold, and the comparison between the rate of change of curvature of low-frequency baseline drift and a preset curvature threshold, the confidence level of the current anomaly attribution is calculated. The preset pulse frequency threshold and preset curvature threshold are set based on historical data and expert experience, and are used to determine whether the degree of high-frequency transient pulses and low-frequency baseline drift reaches the typical characteristics of a certain anomaly state. For example, if the occurrence frequency of high-frequency transient pulses far exceeds the preset pulse frequency threshold, it may be more likely to be attributed to tool chipping or interface impact; if the rate of change of curvature of low-frequency baseline drift is significantly higher than the preset curvature threshold, it may be more likely to be attributed to rapid tool wear or continuous interface degradation. By comprehensively analyzing the results of these subtle characteristics, the confidence level of attributing the current anomaly to a specific mode (such as tool wear or interface micro-degradation) can be calculated more accurately.

[0035] In one embodiment, assuming that during CNC lathe machining, the system identifies a potential anomaly through preliminary anomaly judgment and the construction of a context-related information package. At this point, the system calculates the similarity between this context-related information package and the tool wear pattern in the preset learning sample library to be 0.85, and the similarity to the interface micro-degradation pattern to be 0.82. The difference between the two is 0.03. If the preset discrimination threshold is set to 0.05, since 0.03 is less than 0.05, the system determines that the overall similarity alone is insufficient to clearly distinguish between these two anomalies, thus requiring further analysis. The system then analyzes the sensor signals in the context-related information package, detecting a high-frequency transient pulse occurrence frequency of 120 times / second, while the curvature change rate of the low-frequency baseline drift is 0.03 units / second². The preset pulse frequency threshold is 100 times / second, and the preset curvature threshold is 0.02 units / second². Because the detected pulse frequency (120 pulses / second) is higher than the preset pulse frequency threshold (100 pulses / second), and the rate of curvature change (0.03 units / second²) is higher than the preset curvature threshold (0.02 units / second²), the system comprehensively judges that the current anomaly is more likely caused by rapid tool wear or chipping, rather than microscopic degradation of the interface. Based on this, the system calculates the confidence level of attributing the current anomaly to "tool wear" to be 0.92, which is much higher than the confidence level of attributing it to "interface microscopic degradation." Through this detailed analysis, the system can provide a high-confidence anomaly attribution result, guiding users to take targeted measures such as replacing the tool or adjusting cutting parameters.

[0036] It should be noted that the confidence level of the current anomaly attribution can be understood as a quantitative assessment of the probability or severity of a specific anomaly type (such as tool wear, interface micro-degradation, etc.). This confidence level is the basis for subsequent decisions, and its accuracy directly affects the effectiveness of the avoidance strategy. The pre-defined avoidance strategy library refers to a database containing predefined countermeasures for different anomaly types and confidence levels. These measures may include, but are not limited to, adjusting cutting parameters, replacing tools, stopping for inspection, and adjusting production plans. Selecting multiple candidate avoidance suggestions means choosing several strategies that are most relevant to or most likely to be effective for the current anomaly situation from the avoidance strategy library based on the confidence level of the current anomaly attribution. For example, when the confidence level indicates severe tool wear, the system will select avoidance suggestions related to tool wear. Simulating the impact of candidate avoidance suggestions on short-term production efficiency, tool life, equipment health, and overall production plans refers to using predictive models or simulation technology to evaluate the various consequences that each candidate avoidance suggestion may have in practical applications. Short-term production efficiency refers to the change in output per unit time after taking avoidance measures; tool life refers to the expected service life of tools after taking measures; equipment health refers to the wear, failure risk, etc. of equipment under the influence of avoidance measures; overall production plan refers to the impact of avoidance measures on macro-plans such as production batch delivery time and resource allocation. Through simulation, these impacts can be quantified, resulting in simulation impact score values. Weighted evaluation of multiple simulation impact score values ​​yields a comprehensive benefit score, which means comprehensively considering the scores of each simulation impact result according to preset weights or current production priorities to obtain a single value that comprehensively reflects the merits of avoidance recommendations. For example, in urgent production tasks, the weight of short-term production efficiency may be increased. Selecting the avoidance recommendation with the highest comprehensive benefit score as the final anomaly judgment and alarm generation output means selecting the strategy with the highest comprehensive benefit score from all candidate avoidance recommendations as the final processing solution recommended by the system and outputting it to the user or automation system in the form of an alarm or suggestion.

[0037] In one embodiment, assuming the system determines through the aforementioned steps that there is an anomaly in tool wear during the current CNC lathe machining process, and the confidence level of its attribution to the anomaly is high, the system will then select several candidate avoidance suggestions from a preset avoidance strategy library, such as "reduce cutting speed by 10%", "replace tool", and "adjust feed rate by 5%". Next, the system will simulate and evaluate the impact of these candidate suggestions: For "reduce cutting speed by 10%", the simulation results may show a 5% decrease in short-term production efficiency, a 20% increase in tool life, no significant change in equipment health, and a slight delay in the overall production plan; for "replace tool", the simulation results may show a 15% decrease in short-term production efficiency during the replacement period (downtime), followed by a return to normal and possibly a slight improvement after replacement, with tool life restored to brand new condition and equipment health guaranteed, but resulting in production plan delays due to tool costs and downtime; for "adjust feed rate by 5%", the simulation results may show a 3% decrease in short-term production efficiency, a 10% increase in tool life, no significant change in equipment health, and a minor impact on the production plan. Subsequently, the system weights and evaluates the simulated impact scores based on factors such as the urgency of the current machining task, tool cost, and equipment importance. For example, if the current task is urgent and tool cost is high, the weights of short-term production efficiency and tool life will be increased. After weighted evaluation, a comprehensive benefit score is calculated for each suggestion. Finally, the system selects the avoidance suggestion with the highest comprehensive benefit score (e.g., "replace tool" in an emergency, and "reduce cutting speed by 10%" in a non-emergency situation) as the final anomaly judgment and alarm generation output, and issues specific action instructions to the operator.

[0038] It should be noted that the urgency of the current processing task, the historical equipment failure rate, and the actual tool wear are obtained. Specifically, the urgency of the current processing task can be determined based on the production plan, order priority, or user input; the historical equipment failure rate can be obtained from equipment maintenance records or historical operating data, reflecting the equipment's reliability; the actual tool wear can be obtained in real time through tool life monitoring systems or visual inspection. Based on the urgency, the weight of short-term production efficiency in the overall benefit score calculation is adjusted to obtain the first weight. The higher the urgency, the higher the weight of short-term production efficiency is usually, to prioritize production schedules. Based on the historical equipment failure rate, the weight of equipment health status in the overall benefit score calculation is adjusted to obtain the second weight. Specifically, when the historical equipment failure rate is high, the weight of equipment health status is increased to encourage the system to choose the avoidance strategy with the least impact on equipment health. Based on the actual tool wear, the weight of tool life in the overall benefit score calculation is adjusted to obtain the third weight. In practical applications, when tool wear is significant, the weight of tool life is increased to extend tool life and reduce replacement costs. Based on the urgency, the weight of production batch delivery time in the overall benefit score calculation is adjusted to obtain the fourth weight. For example, in situations with tight delivery deadlines, the weight of production batch delivery time is significantly increased. Multiple simulated impact scores are weighted and evaluated based on a first, second, third, and fourth weight to calculate a comprehensive benefit score. Therefore, by dynamically adjusting the weights of various influencing factors, the actual needs and priorities under the current production environment can be more accurately reflected.

[0039] In one embodiment, suppose a CNC lathe is processing a batch of high-value hardware parts with a tight delivery time. At this point, the system detects an initial anomaly and generates an initial alarm. When performing the final anomaly assessment and alarm generation, the system first determines that the urgency of the current processing task is "high," the historical equipment failure rate is at a "medium" level, and the actual tool wear is "low." Based on this contextual information, the system dynamically adjusts the weights: due to the high urgency, the weights of short-term production efficiency and batch delivery time are significantly increased (e.g., the first and fourth weights are set to 0.4 and 0.3, respectively); the weight of equipment health status remains moderate due to the moderate historical equipment failure rate (e.g., the second weight is set to 0.2); and the weight of tool life is relatively low due to the low tool wear (e.g., the third weight is set to 0.1). Subsequently, the system uses these adjusted weights to perform a weighted evaluation of the simulated impact scores of each candidate avoidance suggestion, calculating a comprehensive benefit score. For example, an avoidance strategy that can quickly restore production but may slightly accelerate tool wear might have a higher overall benefit score, given the current high urgency level, than a strategy that maximizes tool life but significantly extends downtime. In this way, the system can intelligently select the avoidance strategy that best aligns with current production goals and risk appetite, ensuring optimal decision-making at critical moments.

[0040] It should be noted that acquiring historical production efficiency fluctuations, tool wear rates, and equipment micro-vibration patterns refers to the system collecting and storing records of CNC lathe production efficiency changes over time under different machining conditions, tool wear rate data during actual use, and micro-vibration pattern data generated during equipment operation. This historical data provides the foundation for subsequent correlation analysis. Historical production efficiency fluctuations can be understood as the deviation between actual and expected output under different machining tasks and urgency levels; tool wear rate refers to the amount of tool wear per unit time, aiming to quantify tool wear; equipment micro-vibration patterns refer to the subtle vibration characteristics generated by the equipment during operation, which are often closely related to the equipment's health and machining stability. Furthermore, analyzing the correlation between urgency and historical production efficiency fluctuations, tool wear rates, and equipment micro-vibration patterns involves exploring the intrinsic connection between the urgency of the current machining task and the aforementioned historical data through data mining and statistical analysis methods. For example, it can be analyzed whether production efficiency fluctuations, tool wear rates, and equipment micro-vibration patterns exhibit specific trends or patterns under different urgency levels. The aim is to reveal the comprehensive impact mechanism of urgency on production efficiency, tool life, and equipment health. Based on this, and according to the correlation, the potential impact of increased production efficiency on tool life is predicted, yielding the prediction results. This involves constructing a prediction model based on the analyzed correlations to assess the potential impact on tool life when production efficiency is increased to meet urgency requirements. For example, if the analysis indicates that adjustments to certain machining parameters taken to improve efficiency at high urgency levels will significantly accelerate tool wear, the prediction results will reflect this potential negative impact. The aim is to quantify the trade-off between production efficiency and tool life. Finally, based on the prediction results, the weight of short-term production efficiency in the calculation of the overall benefit score is adjusted to obtain the first weight. This refers to dynamically adjusting the importance of short-term production efficiency in the calculation of the overall benefit score based on the above prediction results. For example, if the prediction results show that excessive pursuit of production efficiency at the current urgency level will severely shorten tool life, the weight of short-term production efficiency will be appropriately reduced to avoid losing more than one gains. Conversely, if the prediction results indicate that production efficiency can be improved without significantly affecting tool life, the weight can be appropriately increased. The aim is to achieve a balance between short-term efficiency and long-term benefits.

[0041] In one embodiment, suppose a CNC lathe is processing a batch of high-urgency hardware parts. The system first acquires historical production efficiency fluctuation data for the lathe. For example, under similar urgency levels in the past, a 10% increase in production efficiency typically leads to a 20% increase in tool wear rate. Simultaneously, historical data on tool wear rate and equipment micro-vibration patterns are also acquired. Analysis reveals that under the current urgency level, setting the weight of short-term production efficiency too high, while quickly completing the task, significantly accelerates tool wear and may even lead to premature tool failure, thus affecting subsequent production plans. Based on this correlation analysis, the system predicts the potential negative impact of increased production efficiency on tool life. For example, the prediction shows that increasing the short-term production efficiency weight to 0.4 will shorten tool life by 30%. Based on this prediction, the system intelligently adjusts the weight of short-term production efficiency in the overall benefit score calculation, for example, from the default 0.3 to 0.25, to effectively control tool wear while ensuring a certain increase in production efficiency, thereby achieving the optimal balance between short-term benefits and long-term sustainability.

[0042] It should be noted that obtaining the material type and geometry information of the currently processed hardware part is done automatically or through operator input before performing correlation analysis. This involves acquiring the specific material properties of the hardware part being processed, such as stainless steel, aluminum alloy, and titanium alloy, as well as its geometric features, such as complex surfaces, deep holes, and thin-walled structures. This information is crucial influencing the dynamic behavior of the processing. The preset correlation pattern library is a database storing a large amount of historical processing data and expert experience. This database contains predefined models or patterns relating different hardware part material types and geometries to urgency and production efficiency fluctuations, tool wear acceleration, and equipment micro-vibration patterns. For example, the tool wear acceleration pattern for processing high-hardness materials may differ significantly from that for low-hardness materials; the equipment micro-vibration pattern may be more complex for processing complex geometries. Material-geometry specific correlation patterns refer to the correlation analysis model that is precisely matched or selected from the correlation pattern library based on the currently acquired hardware part material type and geometry information, and best suited to the current processing scenario. For example, when machining thin-walled titanium alloy parts, the system will select a correlation mode specifically for thin-walled titanium alloy structures. This mode may focus more on analyzing the impact of high-frequency vibration on tool life. The purpose is to ensure that subsequent correlation analysis fully considers the uniqueness of the current machining object and avoids deviations that may be caused by using a general model.

[0043] In one embodiment, it is assumed that a CNC lathe is machining a batch of aero-engine blades made of a high-temperature alloy. The geometry is a complex curved thin-walled structure. When analyzing the correlation between urgency and historical production efficiency fluctuations, accelerated tool wear, and equipment micro-vibration patterns, the system first obtains information on the material type of the hardware component as "high-temperature alloy" and the geometric shape as "complex curved thin-walled structure." Subsequently, based on this information, the system retrieves and selects a material-geometry specific correlation pattern from a pre-set correlation pattern library specifically for "machining of complex curved thin-walled structures of high-temperature alloys." This pattern may contain the following characteristics: for high-temperature alloys, the accelerated tool wear pattern may focus more on the high-frequency fluctuations in cutting temperature and cutting force; for complex curved thin-walled structures, the equipment micro-vibration pattern may focus more on resonance phenomena within a specific frequency range and their impact on surface quality and tool life. Based on this specific correlation pattern, the system analyzes how the urgency of the current machining task is related to these specific production efficiency fluctuations, accelerated tool wear, and equipment micro-vibration patterns. For example, at higher urgency levels, even a slight increase in production efficiency may significantly increase the risk of accelerated tool wear in high-temperature alloys or induce unacceptable vibrations during thin-walled structure machining. In this way, this application can provide a highly customized and accurate correlation analysis, thereby providing a more reliable basis for subsequent weight adjustment and avoidance strategy selection.

[0044] It should be noted that obtaining the tool material type, tool coating type, and cutting fluid type used in the current machining process refers to the system automatically or through user input to obtain specific information about the tools used in the current CNC lathe machining task. The tool material type can include cemented carbide, high-speed steel, ceramics, etc.; the tool coating type can include... Cutting fluid types can include emulsions, semi-synthetic cutting fluids, and pure oil cutting fluids. This information is fundamental to understanding tool wear behavior. Based on the tool material type, tool coating type, and cutting fluid type, a corresponding wear mechanism model is selected from a pre-established tool wear mechanism library. This can be understood as the system searching and matching the tool wear mechanism model that best suits the current machining conditions in a pre-built database based on the acquired tool and cutting fluid information. This wear mechanism library stores the main wear mechanisms (e.g., abrasive wear, adhesive wear, diffusion wear, oxidation wear, fatigue wear, etc.) and their mathematical models under different tool-workpiece-cutting fluid combinations. The interaction between urgency and various wear factors in the wear mechanism model is analyzed to obtain the interaction analysis results. Specifically, after selecting a wear mechanism model, the system further evaluates how the urgency of the current machining task (e.g., whether a significant increase in production efficiency is needed) affects the various wear factors in the model. For example, when the urgency is high and the cutting speed and feed rate need to be increased, abrasive wear and diffusion wear may accelerate, while the effect of adhesive wear may be relatively weakened. Furthermore, based on the interaction analysis results, the influence weights of each wear factor on the tool wear rate in the wear mechanism model are adjusted to obtain the adjusted wear mechanism model. This means that, based on the analysis results of the interaction between wear factors and urgency, the parameters of the selected wear mechanism model are adjusted so that the model can more accurately reflect the tool wear behavior under the current specific machining conditions. Therefore, based on the adjusted wear mechanism model and its correlations, the potential impact of increased production efficiency on tool life is predicted, yielding the prediction results. This means that a comprehensive prediction is made by combining the adjusted wear mechanism model, which better reflects actual working conditions, with the previously analyzed correlations between urgency, historical production efficiency fluctuations, tool wear rate, and equipment micro-vibration modes.

[0045] See Figure 2 , Figure 2 This is a schematic diagram of a CNC lathe hardware parts machining process control system provided in one embodiment of this application. The CNC lathe hardware parts machining process control system 200 includes: The acquisition module 210 is used to acquire sensor signals during the machining process of hardware parts on a CNC lathe; The generation module 220 is used to perform preliminary anomaly judgment on sensor signals based on preset judgment rules and generate preliminary alarms; The data recording module 230 is used to receive user feedback on the initial alarm and to acquire sensor signal characteristics and processing parameters related to the initial anomaly judgment. The judgment and generation module 240 is used to judge anomalies and generate alarms based on feedback information, sensor signal characteristics and processing parameters.

[0046] It should be noted that the information interaction and execution process between the above modules are based on the same concept as the method embodiments of this application. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0047] Those skilled in the art will understand that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which can include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media includes, but is not limited to, […]. Flash memory or other memory technologies Digital multifunction discs (DVDs) or other optical disc storage, cartridges, magnetic tapes, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically include computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0048] The foregoing has provided a detailed description of the preferred embodiments of this application. However, this application is not limited to the above-described embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of this application. All such equivalent modifications or substitutions are included within the scope defined in this application.

Claims

1. A method for controlling the machining process of hardware parts on a CNC lathe. Its features are, include: Acquire sensor signals during the machining process of hardware parts on a CNC lathe; Based on preset judgment rules, the sensor signals are initially judged for abnormalities and an initial alarm is generated; Receive user feedback on initial alarms and acquire sensor signal characteristics and processing parameters related to initial anomaly assessment; Based on feedback information, sensor signal characteristics, and processing parameters, anomaly detection and alarm generation are performed.

2. The method according to claim 1, characterized in that, Based on feedback information, sensor signal characteristics, and processing parameters, anomaly detection and alarm generation are performed, including: Obtain background vibration and environmental information of the machine tool related to the initial anomaly assessment; Based on feedback information, sensor signal characteristics, processing parameters, machine tool background vibration, and environmental information, a context-related information package is obtained. When a context-related information packet appears, the similarity between the context-related information packet and the context-related information packets corresponding to different intent tags in the preset learning sample library is calculated. The preset learning sample library is established based on historical anomaly types and historical context-related information packets. Calculate the confidence level of the current anomaly attribution based on similarity; Anomaly detection and alert generation are performed based on the confidence level of the current anomaly attribution.

3. The method according to claim 2, characterized in that, When a context-related information packet appears, the similarity between the context-related information packet and context-related information packets corresponding to different intent tags in a pre-defined learning sample library is calculated, including: Identify features in context-related information packets that are associated with frequency drift or dynamic evolution; By comparing the dynamic evolution features of context-related information packages with the dynamic evolution features of context-related information packages corresponding to different intent labels in the preset learning sample library, the matching results of dynamic evolution features are obtained. Based on the matching results of the dynamic evolution features, the similarity between the new context-related information package and the context-related information packages corresponding to different intent labels in the learning sample library is calculated.

4. The method according to claim 2, characterized in that, The confidence level of the current anomaly attribution is calculated based on similarity, including: Compare the differences between the similarity between the context-related information package and the tool wear pattern, and the similarity between the context-related information package and the interface micro-degradation pattern, and the preset discrimination thresholds. When the difference is less than the preset discrimination threshold, the occurrence frequency of high-frequency transient pulses and the rate of curvature change of low-frequency baseline drift in the analysis context association information packet are analyzed. The confidence level of the current anomaly attribution is calculated based on the comparison between the occurrence frequency of the high-frequency transient pulse and the preset pulse frequency threshold, and the comparison between the curvature change rate of the low-frequency baseline drift and the preset curvature threshold.

5. The method according to claim 1, characterized in that, Anomaly detection and alert generation are performed based on the confidence level of the current anomaly attribution, including: Based on the confidence level of the current anomaly attribution, multiple candidate avoidance suggestions are selected from the preset avoidance strategy library; For each candidate avoidance suggestion, the impact of the candidate avoidance suggestion on short-term production efficiency, tool life, equipment health and overall production plan is simulated to obtain multiple simulation impact result scores; The scores of multiple simulation impact results are weighted and evaluated to calculate the comprehensive benefit score; The avoidance suggestion with the highest overall benefit score is selected as the final anomaly assessment and alarm generation output.

6. The method according to claim 5, characterized in that, The scores of multiple simulated impact results are weighted and evaluated to calculate a comprehensive benefit score, including: Obtain the urgency of the current machining task, the historical equipment failure rate, and the actual wear of the cutting tools; Based on the degree of urgency, the weight of short-term production efficiency in the calculation of the comprehensive benefit score is adjusted to obtain the first weight; Based on the historical failure rate of the equipment, the weight of the equipment health status in the calculation of the comprehensive benefit score is adjusted to obtain the second weight; Based on the actual wear of the tool, the weight of tool life in the calculation of the comprehensive benefit score is adjusted to obtain the third weight; Based on the level of urgency, the weight of production batch delivery time in the calculation of the overall benefit score is adjusted to obtain the fourth weight; The comprehensive benefit score is calculated by weighting the scores of multiple simulated impact results according to the first weight, the second weight, the third weight, and the fourth weight.

7. The method according to claim 6, characterized in that, Based on the level of urgency, the weight of short-term production efficiency in the calculation of the overall benefit score is adjusted to obtain the first weight, which includes: Acquire historical production efficiency fluctuations, tool wear rates, and equipment micro-vibration patterns; Analyze the correlation between the level of urgency and historical production efficiency fluctuations, tool wear rate, and equipment micro-vibration modes; Based on the correlation, the potential impact of improved production efficiency on tool life is predicted, and the prediction results are obtained. Based on the forecast results, the weight of short-term production efficiency in the calculation of the comprehensive benefit score is adjusted to obtain the first weight.

8. The method according to claim 7, characterized in that, The correlation between urgency level and historical production efficiency fluctuations, accelerated tool wear, and equipment micro-vibration modes was analyzed, including: Obtain information on the material type and geometry of the hardware parts currently being processed; Based on the material type and geometric shape information of the hardware parts, select the corresponding material-geometry specific association mode from the preset association mode library; Based on the material-geometry specific correlation model, the correlation between urgency and historical production efficiency fluctuations, tool wear acceleration, and equipment micro-vibration modes is analyzed.

9. The method according to claim 7, characterized in that, Based on the correlation, the potential impact of increased production efficiency on tool life is predicted, yielding the following prediction results: Obtain the type of tool material, tool coating type, and cutting fluid used in the current machining process; Based on the tool material type, tool coating type, and cutting fluid type, select the corresponding wear mechanism model from the preset tool wear mechanism library; The interaction between the urgency level and various wear factors in the wear mechanism model was analyzed, and the results of the interaction analysis were obtained. Based on the interaction analysis results, the influence weights of each wear factor on the tool wear rate in the wear mechanism model were adjusted to obtain the adjusted wear mechanism model. Based on the adjusted wear mechanism model and correlation, the potential impact of improved production efficiency on tool life is predicted, and the prediction results are obtained.

10. A control system for machining hardware parts on a CNC lathe, characterized in that, include: The acquisition module is used to acquire sensor signals during the machining process of hardware parts on a CNC lathe; The generation module is used to perform preliminary anomaly judgment on sensor signals based on preset judgment rules and generate preliminary alarms; The data logging module is used to receive user feedback on the initial alarm and to acquire sensor signal characteristics and processing parameters related to the initial anomaly judgment. The judgment and generation module is used to judge anomalies and generate alarms based on feedback information, sensor signal characteristics, and processing parameters.

Images