A numerical control boring and milling tool cutting force intelligent monitoring method and system

Abstract

The present application belongs to the field of numerical control machine tool processing technology, and particularly relates to a kind of intelligent monitoring method and system of tool cutting force of numerical control boring and milling, and the method comprises: obtaining the real-time main force, real-time transverse force, real-time longitudinal force sequence in numerical control boring and milling equipment processing task, real-time spindle speed, signal sampling frequency, basic statistical threshold;Obtain characteristic force;Calculate the historical mean of current time, recorded as historical mean;Calculate current historical deviation, determine the sliding window width;Calculate instantaneous deviation;Obtain the change amplitude, change trend of characteristic force of current time, the severity of deviation from its historical normal, the maximum value in both, get risk index;According to the basic statistical threshold, compare whether the risk index is abnormal, generate processing process digital archives.The present application solves the technical problems of high false positive rate and poor reliability caused by the inability to adapt to normal tool wear in the prior art.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of numerical control machine tool processing. More particularly, the present application relates to a kind of intelligent monitoring method and system of tool cutting force of numerical control boring and milling. BACKGROUND

[0002] In modern manufacturing, numerical control boring and milling processing is a key link in producing high-precision, high-value parts, especially in machining large, heavy or high-value workpieces, such as complex structural parts in the aerospace field or key components of mining machinery, the machining precision and surface quality are crucial. In the machining process, the state of the tool is directly related to the final product quality and production efficiency, in order to prevent the tool from accidentally collapsing, breaking and damaging expensive workpieces or machine tool spindles, and to effectively manage the wear state of the tool, the industry generally uses online monitoring technology based on cutting force signals.

[0003] The existing cutting force monitoring method usually collects cutting force signals in real time through force sensors installed on the machine tool, and when some characteristics of the signals, such as amplitude, exceed a pre-set fixed threshold, the system triggers an alarm or stops. However, this method relying on a static threshold has an inherent defect that is difficult to overcome when facing long-term, continuous machining tasks, that is, the tool will inevitably wear slowly during normal use. This normal wear will cause the baseline level of cutting force to rise slowly and continuously. Since the fixed alarm threshold cannot adapt to the change of cutting force data caused by normal wear, the system will frequently trigger false alarms when the tool is still in a healthy and usable state.

[0004] This frequent false alarm phenomenon can seriously interfere with normal production rhythm and cause operators to lose trust in the monitoring system, eventually, the operator may choose to manually adjust the alarm threshold to an unreasonable level or even turn off the monitoring system, which makes the system useless when a tool collapse or other accidents occur, completely losing its intended protection. SUMMARY

[0005] To solve the technical problems of high false alarm rate and poor reliability caused by the inability of the existing technology to adapt to normal tool wear, the present application provides solutions in the following aspects.

[0006] In a first aspect, the present application provides an intelligent monitoring method of tool cutting force of numerical control boring and milling, comprising:

[0007] Obtain the real-time main force, real-time lateral force, real-time longitudinal force sequence in the machining task of the numerical control boring and milling equipment, real-time spindle speed, signal sampling frequency, basic statistical threshold; and according to the relative size relationship between the resultant force of the real-time main force and the real-time lateral force, the real-time longitudinal force, the characteristic force is obtained; according to the characteristic force at the current time, the historical mean value at the last time, the historical deviation, the historical mean value at the current time is calculated through the deviation adaptive weighting mode, and the historical mean value is recorded; according to the characteristic force at the current time, the historical mean value and the historical deviation at the last time, the current historical deviation is calculated through the incremental updating mode, and the sliding window width is determined according to the real-time spindle speed and the signal sampling frequency; according to the absolute deviation of the characteristic force of each sampling point in the sliding window at the current time and the mean value of the characteristic force in the window, the instantaneous deviation is calculated; the change amplitude of the characteristic force at the current time, the severity of the change trend deviating from its historical normality, and the maximum value of the two are obtained, and the risk index is obtained; according to the basic statistical threshold, whether the risk index is abnormal is compared, and after the machining is finished, the digital archive of the machining process is generated.

[0008] The present application solves the three core problems of traditional monitoring through multi-dimensional data acquisition, dynamic model construction and intelligent analysis: first, guarantee the data reliability, with clear initial conditions and high-frequency acquisition mechanism, so that the system can run stably from the beginning of machining, avoiding errors caused by data missing or fuzzy calculation starting point; second, strengthen the fault recognition ability, through the characteristic force design to amplify the abnormality of real-time lateral force and real-time longitudinal force, combined with deviation adaptive weighting and instantaneous deviation calculation, it can not only track the slow change of normal tool wear, but also quickly respond to sudden failure, greatly improve the sensitivity and recognition speed of various abnormalities, reduce false alarm and missed report; third, accelerate the process optimization, the digital archive of the machining process records the risk index, force signal and machine tool state, provides clear reference for engineers to diagnose faults and improve processes, and optimizes tool path or cutting parameters through analyzing the change of force signal before and after alarm, to reduce the failure rate from the root.

[0009] Preferably, the real-time main force, real-time lateral force, real-time longitudinal force sequence in the machining task of the numerical control boring and milling equipment, real-time spindle speed, signal sampling frequency, basic statistical threshold are obtained, including:

[0010] In a predetermined time period, the data sequence including the real-time main force sequence, the real-time lateral force sequence and the real-time longitudinal force sequence is obtained by the sensor and the acquisition unit on the machine tool in a high-frequency sampling manner; the set speed of the current spindle is extracted from the numerical control system of the machine tool in real time, and is recorded as the real-time spindle speed; the preset signal sampling frequency is read from the configuration of the sensor and the acquisition unit, and is recorded as the signal sampling frequency; the constant threshold based on statistical principles for final judgment is extracted from the configuration of the risk index generation and decision module, and is recorded as the basic statistical threshold.

[0011] Preferably, the characteristic force satisfies the following expression:

[0012] ;

[0013] In the formula, represents the characteristic force at the current time within a preset time period; represents the real-time main force at the current time within a preset time period; and represents the real-time lateral force and the real-time longitudinal force at the current time; represents an exponential function with a natural constant as the base; represents an absolute value function; represents a very small positive number to ensure that the denominator is not zero.

[0014] When monitoring the tool state, the application can focus on the changes of forces that affect the stability of machining. When these forces abnormally increase greatly, the changes can be more obviously reflected, so that even if the tool initially has a small problem, it can be detected in time, and an alarm will not be given until the problem is serious enough to cause workpiece damage or equipment failure. The application effectively reduces the loss caused by tool problems, makes the monitoring more sensitive, can prevent risks in advance, and ensures that workers can intervene and handle the tool abnormalities in time during the machining process.

[0015] Preferably, the historical mean satisfies the following expression:

[0016] ;

[0017] In the formula, represents the historical mean at the current time; represents the historical mean at the previous time, representing the long-term average level of history; represents the characteristic force at the current time; represents the historical deviation at the previous time; represents a hyperbolic tangent function, and the output value range is (-1, 1); represents a very small positive number to ensure that the denominator is not zero.

[0018] The application updates the historical mean by using a deviation adaptive weighting strategy, so that the system can dynamically track the slow changes of normal tool wear and quickly respond to sudden abnormalities. This method avoids false alarms of the traditional fixed parameter model for normal wear, maintains smooth tracking of long-term trends, and enhances the sensitivity to short-term mutations. For example, when the tool gradually increases in force value due to normal wear, the system will not misjudge as abnormal, but when a sudden collapse of the blade occurs, the system can quickly capture and trigger an early warning, thereby improving the stability and adaptability of the monitoring system.

[0019] Preferably, the current historical deviation is calculated, including:

[0020] Based on the historical deviation of the last time, the characteristic force of the current time, and the historical mean value of the current time, the difference between the distance of the characteristic force of the current time deviating from its historical mean value and the historical deviation of the last time is calculated, which is recorded as the unexpected degree; the maximum value of the positive integer time sequence number of the current time in time sequence is taken within a preset time period, which is recorded as the maximum time sequence number; the unexpected degree is compared with the maximum time sequence number to calculate the current historical deviation.

[0021] Preferably, the sliding window width is determined, comprising:

[0022] The coefficient for covering n main shaft rotation periods, n≥1, is obtained, which is recorded as the main shaft rotation period coefficient; the main shaft revolutions per second is obtained by dividing the real-time main shaft speed by 60; the signal sampling frequency is compared with the main shaft revolutions per second, and the obtained result is multiplied by the main shaft rotation period coefficient to obtain the sliding window width.

[0023] Preferably, the instantaneous deviation satisfies the following expression:

[0024] ;

[0025] In the formula, The instantaneous deviation of the current time is represented; The sliding window width is represented; j is the jth sampling point in the sliding window; The characteristic force of the jth sampling point in the sliding window is represented; The mean value of all in the sliding window is represented.

[0026] The present application accurately locates the force signal fluctuation in a very short time by calculating the instantaneous deviation in the sliding window, and is sensitive to sudden impact or vibration anomaly. This method makes up for the defects of insufficient response to instantaneous abnormality of traditional global analysis. For example, when the tool suddenly vibrates violently in a few rotation periods, the instantaneous deviation index will quickly rise, even if the global mean value does not change significantly, the system can also timely warn, and the identification speed and accuracy of the precursor of malignant failure are improved.

[0027] Preferably, the risk index satisfies the following expression:

[0028] ;

[0029] In the formula, The risk index of the current time is represented; The historical mean value of the current time is represented; The characteristic force of the current time is represented; The historical deviation of the current time is represented; The instantaneous deviation of the current time is represented; represents the maximum value function; represents a minimum positive number, which guarantees the denominator not to be 0.

[0030] The risk index in the application is a two-dimensional evaluation index integrating amplitude and fluctuation, and comprehensively covers different failure modes. No matter whether the tool failure is represented by amplitude mutation or fluctuation increase, the system can respond sensitively by taking the maximum value. For example, when tool wear causes gradual increase of fluctuation but the amplitude is not out of limit, the fluctuation anomaly part will trigger the risk index to rise; and when sudden collapse of blade causes sudden increase of amplitude, the amplitude anomaly part will dominate the risk index change. This design ensures comprehensive coverage and accurate identification of various failures by the monitoring system.

[0031] Preferably, according to the basic statistical threshold, it is compared whether the risk index is abnormal, and after the machining is finished, a digital archive of the machining process is generated, including:

[0032] The risk index at the current time is compared with the basic statistical threshold to determine whether cutting abnormality occurs; when a complete machining task is finished, whether it is completed normally or interrupted due to alarm, the system binds the data of the whole machining task with the unique identification code of the workpiece to generate a digital archive of the machining process. The digital archive of the machining process includes: a risk index sequence; associated data when the risk index exceeds the basic statistical threshold; real-time main force, real-time lateral force and real-time longitudinal force data within 2 seconds before and after the alarm time; machine tool coordinates and spindle speed at the alarm time.

[0033] In a second aspect, the application provides a tool cutting force intelligent monitoring system for numerical control boring and milling, which comprises a processor and a memory, and the memory stores computer program instructions.

[0034] By using the above technical solution, the above-mentioned tool cutting force intelligent monitoring method for numerical control boring and milling is generated into a computer program and stored in the memory to be loaded and executed by the processor, so that a terminal device is made according to the memory and the processor, and the use is convenient.

[0035] The beneficial effects of the present application are that the technical scheme of the present application creates three values for the numerical control boring and milling processing industry: in terms of production efficiency, real-time monitoring and intelligent early warning mechanism reduces unplanned downtime caused by tool failure, such as the system can identify the precursor of tool blade collapse in advance and replace the tool in time to avoid workpiece scrap and equipment damage, effectively guaranteeing the continuity of processing and the utilization rate of production capacity; in terms of cost control, without relying on expensive external sensors, high-precision monitoring can be achieved with the help of machine tool inherent signals and algorithm optimization, reducing the cost of equipment modification, and data-driven process optimization strategy reduces tool consumption and scrap rate, further reducing production cost; in terms of quality and competitiveness, accurate tool state monitoring ensures stable processing, improves workpiece precision and surface quality, and provides technical support for enterprises to undertake high-value-added orders, and the accumulation of digital archives helps enterprises establish a process knowledge base, form a differentiated technical advantage, and enhance market competitiveness. BRIEF DESCRIPTION OF DRAWINGS

[0036] Figure 1 is a flow chart schematically showing a tool cutting force intelligent monitoring method of a numerical control boring and milling machine in the present application. DETAILED DESCRIPTION

[0037] The embodiment of the present application discloses a tool cutting force intelligent monitoring method of a numerical control boring and milling machine, referring to Figure 1 , comprising steps S1-S4:

[0038] S1: Obtain the real-time main force, real-time lateral force, real-time longitudinal force sequence in the numerical control boring and milling equipment processing task, real-time spindle speed, signal sampling frequency, and basic statistical threshold value; and obtain the characteristic force according to the relative size relationship between the resultant force of the real-time main force and the real-time lateral force and the real-time longitudinal force.

[0039] It should be noted that in order to enable the monitoring system to accurately perceive the actual state of the numerical control boring and milling processing process and make intelligent analysis and decision based thereon, the system needs to obtain a series of basic real-time data from the processing equipment, which are the direct input for building all subsequent intelligent algorithm models, including cutting force information reflecting the interaction between the tool and the workpiece, and parameters such as spindle speed and sampling frequency. At the same time, considering that the present application adopts multiple recursive calculation methods, these methods need a clear starting point in mathematics to run smoothly. If the initial conditions of the calculation are not clearly defined, the algorithm will not be able to obtain effective input or reference values at the first moment of processing, resulting in the entire monitoring system being unable to start or producing calculation errors. Therefore, the present application obtains clear initial data types and their time dimensions, and sets strict algorithm initial conditions.

[0040] Specifically, a single, continuous machining process being performed on a specific CNC boring and milling equipment is taken as a target machining task; three-directional cutting forces in the target machining task process are acquired in a high-frequency sampling manner through sensors and a collection unit on the machine tool within a preset time period to form discrete time sequences including a real-time main force sequence, a real-time lateral force sequence and a real-time longitudinal force sequence; a set speed of a current spindle is extracted from a numerical control system of the machine tool and recorded as a real-time spindle speed; a preset signal sampling frequency is read from a configuration of the sensors and the collection unit and recorded as a signal sampling frequency; and a constant threshold based on a statistical principle for final judgment is extracted from a configuration of a risk index generation and decision module and recorded as a basic statistical threshold.

[0041] It should be noted that, in order to ensure the smooth start of the recursive calculation, the application defines the initial conditions: at the first time point of the preset time period, the initial characteristic force is set, the initial historical mean is set, and the initial historical deviation is set. All subsequent calculations start at the second time point of the preset time period.

[0042] At this point, the real-time main force sequence, the real-time lateral force sequence, the real-time longitudinal force sequence, the real-time spindle speed, the signal sampling frequency and the basic statistical threshold of the target machining task are obtained.

[0043] It should be noted that normal wear of the tool or impending failure such as tool breakage is usually not only manifested as a simple increase in the total cutting force, but also in the lateral force component, i.e., the real-time lateral force and the real-time longitudinal force, which will appear a disproportionate and dramatic increase relative to the main cutting force. The subtle but critical precursor information of the traditional cutting force monitoring method, such as simple vector synthesis of three-directional forces, is easily submerged, resulting in the system being unable to identify the real abnormal state of the tool in time. Therefore, in order to improve the sensitivity and early warning capability of the monitoring system to the deterioration of the tool state, it is necessary to construct a comprehensive index that can accurately capture and amplify the abnormal growth of the lateral force, which is recorded as the characteristic force. The characteristic force not only reflects the overall cutting load, but also shows a high sensitivity to the imbalance between the lateral force and the real-time main force, so as to provide a clear early warning signal before the tool really fails, avoiding damage to expensive workpieces and unplanned interruptions of production.

[0044] Preferably, the real-time main force sequence, the real-time lateral force sequence and the real-time longitudinal force sequence are acquired; a very small positive number is set; the characteristic force at the current time within the preset time period is calculated and recorded as

[0045] The characteristic force satisfies the following expression:

[0046] ;

[0047] wherein, represents the characteristic force at the current time within the preset time period; represents the real-time main force at the current time within the preset time period; and represents the real-time lateral force and the real-time longitudinal force at the current time; represents an exponential function with a natural constant as the base; represents an absolute value function; represents a very small positive number to ensure that the denominator is not zero.

[0048] wherein, the fractional term A force imbalance ratio is constructed, which has a physical meaning of evaluating the difference between the resultant force of the real-time lateral force and the real-time longitudinal force and the real-time main force, reflecting whether the force distribution in the cutting process is balanced; The function acts as a nonlinear amplifier to exponentially amplify the force imbalance ratio, and when the real-time lateral force and the real-time longitudinal force increase disproportionately, the amplification effect far exceeds the linear relationship; In the formula, the basic real-time main force is multiplied by the exponential amplification result, so that the final characteristic force not only contains the basic size of the cutting load, but also integrates an amplification factor that is extremely sensitive to the deterioration of the tool state such as wear and edge collapse. When the tool state deteriorates and causes the lateral force to increase, the characteristic force will increase dramatically with an amplitude far exceeding the original force signal.

[0049] Preferably, all characteristic forces within the preset time are calculated to generate a characteristic force sequence.

[0050] S2: According to the characteristic force at the current time, the historical mean value at the last time, and the historical deviation, the historical mean value at the current time is calculated by a deviation self-adaptive weighting method, which is recorded as historical mean value.

[0051] It should be noted that in the long continuous machining process of numerical control boring and milling, the normal wear of the tool will cause the cutting force to show a slow and progressive growth trend, and if the historical mean model only uses the traditional simple arithmetic mean or the fixed parameter moving average, it will not be able to intelligently distinguish between this normal trend change and the sudden fault anomaly. For example, a fixed parameter average model may mistakenly identify the slow rise in force value caused by normal wear as an anomaly, thereby triggering frequent false alarms. In order to enable the monitoring system to establish a truly dynamic benchmark, this benchmark not only needs to smoothly track the normal wear process of the tool, but also needs to have the flexibility to quickly respond to sudden changes, and this response speed cannot be fixed in advance, but should be self-adjusted according to the unexpected degree of the current signal deviation from the historical state. Therefore, the present application adopts a deviation self-adaptive weighting strategy to update the historical mean, which is the key to solving the problem of insufficient stability of traditional mean models in the face of complex working conditions, allowing the system to remain stable during smooth periods and quickly adapt during dramatic changes.

[0052] Specifically, the feature force at the current time in the feature force sequence is obtained; the historical mean of the feature force at the previous time is obtained by deviation self-adaptive weighting update, denoted as historical mean; the historical deviation of the feature force at the previous time is obtained by increment smoothing iteration, denoted as historical deviation; a minimum positive number is set The historical mean at the current time is calculated by a two-step method, denoted as historical mean, including:

[0053] The historical mean satisfies the following expression:

[0054] ;

[0055] In the formula, historical mean at the current time; historical mean at the previous time, representing the long-term average level of history; feature force at the current time; historical deviation at the previous time; denotes the hyperbolic tangent function, and its output value range is (-1, 1); denotes a minimum positive number, which ensures that the denominator is not 0.

[0056] In the formula, The calculation of first normalizes the absolute deviation of the current feature force and the historical mean at the previous time by the historical deviation at the previous time, and the ratio reflects the severity of the current signal deviating from its historical normal state, and then the severity of the current signal deviating from its historical normal state is smoothly mapped to a weight value greater than or equal to 0 and less than 1 by the function; The calculation of is a dynamic weighted sum, and the weight The size determines the degree of trust of the system to new data and old history; An intelligent, nonlinear smoothing filter is implemented, when the signal changes dramatically, Increasing and approaching to 1, the system quickly adopts new signals to track mutations, when the signal is stable, Decreasing and approaching to 0, the system more retains the historical mean to maintain stability.

[0057] S3: According to the characteristic force of the current time, the historical mean and the historical deviation of the last time, the current historical deviation is calculated by incremental updating, and the sliding window width is determined according to the real-time main shaft speed and the signal sampling frequency; the instantaneous deviation is calculated according to the absolute deviation of the characteristic force of each sampling point in the sliding window of the current time and the mean value of the characteristic force in the window.

[0058] The stability of the numerical control boring and milling process is not constant, even under the tool health state, the cutting force signal will exist a certain degree of fluctuation, in order to accurately judge whether the current signal fluctuation is abnormal, the system needs a benchmark that can dynamically reflect the average fluctuation amplitude in the whole processing history. If only the static fluctuation index is used, when the processing conditions such as material batch and tool wear stage change slightly, the original static index may no longer be applicable, resulting in misjudgment of normal fluctuation. Therefore, the incremental updating method is used to calculate the historical deviation, in order to establish a global fluctuation index that can continuously learn and correct with the processing process. This index can dynamically capture the average dispersion degree of the fluctuation amplitude of the cutting force in the processing process, provide a reliable evaluation standard for the realization of the subsequent risk index normalization, and ensure that no matter what stage the processing is in, the system can evaluate the relative severity of the current fluctuation based on the latest historical data.

[0059] Specifically, based on the historical deviation of the last time, the characteristic force of the current time, the historical mean of the current time, the historical deviation of the current time is calculated by incremental updating, which is recorded as the current historical deviation, including:

[0060] The current historical deviation satisfies the following expression:

[0061] ;

[0062] In the formula, The current historical deviation is represented by D; The historical mean of the current time is represented by M; The characteristic force of the current time is represented by F; The historical deviation of the last time is represented by D; The distance of the characteristic force of the current time deviating from its historical mean is represented by F; Ensure that the denominator is the minimum of 1, to avoid the denominator of zero from the procedure.

[0063] In the formula, An incremental correction, the numerator is the distance of the current time of the characteristic force deviates from its historical mean and the difference of the last time of the historical deviation, represents the unexpected degree of new information, the denominator as an online learning rate of time-decreasing, ensures that with the processing time, the influence of the individual time of the unexpected degree on the overall historical deviation is smaller and smaller; Add the last time of the historical deviation and the incremental correction to realize the online learning and continuous correction of the historical deviation, so that the current historical deviation can stably converge to the arithmetic mean of all historical absolute deviations.

[0064] It should be noted that the cutting force signal of numerical control boring and milling processing has periodic characteristics, and this periodicity is closely related to the rotation of the spindle. Each cutting edge of the tool will interact with the workpiece once or more times in a rotation, generating a force signal with a specific frequency and shape. If the width of the sliding window cannot accurately match this physical periodicity, the data in the window will not fully reflect the dynamic characteristics of a single or multiple complete cutting edge periods when calculating the local volatility, which may lead to inaccurate capture of local anomalies by the system or introduce irrelevant noise. Therefore, the sliding window width is combined with the real-time spindle speed and signal sampling frequency to ensure that the window covers one or more tool rotation periods, which is the key to accurately capturing short-term and local cutting dynamics, so that the calculation of instantaneous deviation can truly reflect the real performance of the tool in a single or a few rotation periods.

[0065] Preferably, the coefficient for covering n spindle rotation periods is obtained, n≥1, denoted as the spindle rotation period coefficient, the real-time spindle speed is divided by 60 to obtain the number of spindle revolutions per second, the signal sampling frequency is compared with the number of spindle revolutions per second, and the obtained result is multiplied by the spindle rotation period coefficient to obtain the sliding window width.

[0066] It should be noted that in addition to long-term average fluctuation trend, sudden impact or severe vibration of the tool in extremely short time, such as one or several spindle rotation periods, is often a direct precursor of catastrophic failure such as chipping and chatter, and if only relying on the historical deviation of global characteristic force, these instantaneous and local severe fluctuations may be ignored, resulting in the system failing to timely warn. Therefore, in order to improve the sensitivity and response speed of the system to instantaneous abnormalities, an index capable of accurately evaluating the dispersion degree of cutting force in the current extremely short time window is needed. The calculation of instantaneous deviation is to analyze the local physical interaction between the tool and the workpiece in the time scale, which can effectively locate the local dynamic abnormalities with short duration but strong destructive power by calculating the average absolute deviation in a sliding window synchronized with the tool rotation period.

[0067] Preferably, the average absolute deviation of the characteristic force sequence in the sliding window before the current time is calculated, denoted as the instantaneous deviation, including:

[0068] The instantaneous deviation satisfies the following expression:

[0069] ;

[0070] In the formula, represents the instantaneous deviation at the current time; represents the width of the sliding window; j is the jth sampling point in the sliding window; represents the characteristic force of the jth sampling point in the sliding window; is the mean value of all in the sliding window.

[0071] In the formula, the denominator represents the width of the sliding window, i.e. the number of samples in the sliding window, ensuring that the calculation result is an average value; The sum of the absolute deviations of all characteristic force points in the sliding window from the mean value of the characteristic forces of all sampling points in the window is calculated, reflecting the dispersion degree of the data in the window; By calculating the average absolute deviation in a sliding window, the evaluation of instantaneous volatility is realized, and becomes a very sensitive index to short-term process stability.

[0072] S4: Obtain the change amplitude of the characteristic force at the current time, the severity of the deviation of the change trend from its historical normal state, and take the maximum value of the two to obtain the risk index; compare whether the risk index is abnormal according to the basic statistical threshold, and generate a digital archive of the machining process after the machining is completed.

[0073] The historical mean value, the historical deviation and the instantaneous deviation are used to calculate a normalized amplitude abnormality score and a normalized fluctuation abnormality score respectively, and the maximum of the two is taken to obtain a risk index; the risk index is compared with a basic statistical threshold to determine whether a cutting abnormality occurs.

[0074] It should be noted that the occurrence of tool failure is various, and not all failures are simply manifested as a sudden increase in cutting force amplitude. For example, some tool wear may cause an increase in cutting force fluctuation, while the average amplitude change is not obvious, and some sudden flank collapse events may be accompanied by a huge force impact and severe vibration. If the risk assessment model only focuses on a single dimension of abnormality, such as only amplitude exceeding, it may miss those failure modes mainly with a sharp increase in fluctuation, thereby reducing the comprehensiveness and reliability of the monitoring system. Therefore, in order to build an index that can comprehensively reflect the actual risk of the tool, a model that can evaluate the abnormality degree of the signal in two independent dimensions of amplitude deviation and sharp increase in fluctuation and intelligently integrate these information is needed. This model ensures that no matter how the failure is manifested, the monitoring system can capture and judge from multiple angles, improving the accuracy and coverage of failure identification.

[0075] Specifically, the characteristic force at the current time, the historical mean value at the current time, the historical deviation at the current time, and the instantaneous deviation are obtained; a minimum positive number is set; a risk index of abnormality of the tool cutting force at the current time is calculated, denoted as risk index, including:

[0076] The risk index satisfies the following expression:

[0077] ;

[0078] In the formula, represents the risk index at the current time; represents the historical mean value at the current time; represents the characteristic force at the current time; represents the historical deviation at the current time; represents the instantaneous deviation at the current time; represents the maximum function; represents a minimum positive number, which ensures that the denominator is not 0.

[0079] In the formula, represents the amplitude abnormality score, which normalizes the absolute difference between the characteristic force at the current time and the historical mean value by the historical average deviation, and evaluates the severity of the change in amplitude of the characteristic force at the current time deviating from its historical normality; The abnormal fluctuation anomaly represents the instantaneous fluctuation of the characteristic force at the current time, which is normalized by the historical average fluctuation, and evaluates the severity of the trend of the characteristic force at the current time deviating from its historical normality; By taking the maximum value, a logical judgment about or is realized, ensuring that the risk index will increase as long as one of them shows an anomaly, whether due to amplitude mutation or due to a sharp increase in volatility, thereby responding sensitively to different types of failure modes.

[0080] It should be noted that for enterprises focusing on the production of large, high-value heavy equipment parts, the monitoring system of numerical control boring and milling processing is far from enough to achieve alarm shutdown. Each failure shutdown or processing completion brings valuable process information and quality data. If this information cannot be systematically recorded, archived and analyzed, the enterprise will not be able to deeply understand the root cause of the failure, nor can it provide objective evidence for future product quality traceability, nor can it accumulate and optimize manufacturing process knowledge. Therefore, in order to enable the monitoring system to evolve from a passive safety tool to an active process improvement and quality management platform, the present application also includes a complete data archiving and diagnostic analysis mechanism, which not only provides engineers with powerful fault analysis tools, but also records the digital characteristics of the processing process.

[0081] Preferably, the risk index at the current time is compared with the basic statistical threshold to determine whether a cutting anomaly occurs; when a complete processing task is completed, whether it is completed normally or interrupted due to an alarm, the system binds the data of the processing task with the unique identification code of the workpiece to generate a processing process digital archive, which includes: a risk index sequence; associated data when the risk index exceeds the basic statistical threshold; real-time main force, real-time lateral force, and real-time longitudinal force data within 2 seconds before and after the alarm time; machine tool coordinates and spindle speed at the alarm time.

[0082] The above system also includes other components such as communication bus and communication interface, which are well known to those skilled in the art, and their settings and functions are known in the art, so they will not be described here.

[0083] Although the present specification has shown and described multiple embodiments of the present application, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will think of many changes, modifications and alternatives without departing from the spirit and principles of the present application. It should be understood that various alternatives to the embodiments of the present application described herein can be employed in practicing the present application.​

Claims

1. A method for intelligent monitoring of cutting force of CNC boring and milling tools, characterized in that, include: Acquire real-time principal force, real-time transverse force, real-time longitudinal force sequence, real-time spindle speed, signal sampling frequency, and basic statistical thresholds in the machining task of CNC boring and milling equipment; The characteristic force is obtained based on the relative magnitude of the resultant forces of the real-time principal force, the real-time lateral force, and the real-time longitudinal force. Characteristic forces are In the formula, This represents the characteristic force at the current moment within a preset time period; This represents the real-time principal force at the current moment within a preset time period; and This indicates the real-time lateral force and real-time longitudinal force at the current moment; Represents an exponential function with the natural constant as the base; Represents the absolute value function; It represents a very small positive number, used to ensure that the denominator is not zero; Based on the characteristic force at the current moment, the historical mean at the previous moment, and the historical deviation, the historical mean at the current moment is calculated using an adaptive weighting method based on the deviation, and is denoted as the historical mean. Based on the characteristic force at the current moment, the historical average, and the historical deviation at the previous moment, the current historical deviation is calculated through incremental updates, and the width of the sliding window is determined based on the real-time spindle speed and signal sampling frequency; the instantaneous deviation is calculated based on the absolute deviation between the characteristic force of each sampling point within the sliding window at the current moment and the average characteristic force within the window. The system obtains the magnitude of the change in the characteristic force at the current moment and the severity of the deviation of the change trend from its historical normal. The maximum value of the two is taken to obtain the risk index. Based on the basic statistical threshold, the system compares whether the risk index is abnormal. After the processing is completed, a digital archive of the processing process is generated.

2. The intelligent monitoring method for cutting force of CNC boring and milling tools according to claim 1, characterized in that, The acquisition of real-time principal force, real-time transverse force, real-time longitudinal force sequence, real-time spindle speed, signal sampling frequency, and basic statistical thresholds in the CNC boring and milling machine machining task includes: Within a preset time period, data sequences including real-time principal force sequence, real-time transverse force sequence, and real-time longitudinal force sequence are acquired through high-frequency sampling via sensors and acquisition units on the machine tool; the current spindle speed is extracted in real-time from the CNC system of the machine tool and recorded as the real-time spindle speed; the preset signal sampling frequency is read from the configuration of the sensors and acquisition units and recorded as the signal sampling frequency; and a constant threshold based on statistical principles for final judgment is extracted from the configuration of the risk index generation and decision-making module and recorded as the basic statistical threshold.

3. The intelligent monitoring method for cutting force of CNC boring and milling tools according to claim 1, characterized in that, The historical mean satisfies the following expression: ； In the formula, This represents the historical average at the current moment; It represents the historical average at the previous moment, and represents the long-term average level in history; Represents the characteristic force at the current moment; Indicates the historical deviation from the previous moment; Let represent the hyperbolic tangent function, whose output range is (-1, 1); It represents a very small positive number, and guarantees that the denominator is not 0.

4. The intelligent monitoring method for cutting force of CNC boring and milling tools according to claim 1, characterized in that, The calculation of the current historical deviation includes: Based on the historical deviation of the previous moment, the characteristic force of the current moment, and the historical mean of the current moment, calculate the difference between the distance of the characteristic force of the current moment from its historical mean and the historical deviation of the previous moment, and record it as the degree of surprise; take the maximum value of the positive integer time sequence number of the current moment, which is counted from 1 in time sequence within the preset time period, and record it as the maximum time sequence number; compare the degree of surprise with the maximum time sequence number to calculate the current historical deviation.

5. The intelligent monitoring method for cutting force of CNC boring and milling tools according to claim 1, characterized in that, Determining the width of the sliding window includes: Obtain the coefficients used to cover n spindle rotation cycles, where n≥1, and denote them as spindle rotation cycle coefficients. Divide the real-time spindle speed by 60 to obtain the spindle speed per second. Compare the signal sampling frequency with the spindle speed per second, and multiply the result by the spindle rotation cycle coefficients to obtain the sliding window width.

6. The intelligent monitoring method for cutting force of CNC boring and milling tools according to claim 1, characterized in that, The calculated instantaneous deviation satisfies the following expression: ； In the formula, Indicates the instantaneous deviation at the current moment; Indicates the width of the sliding window; j represents the j-th sampling point within the sliding window; This represents the characteristic force of the j-th sampling point within the sliding window; For all within this sliding window The mean.

7. The intelligent monitoring method for cutting force of CNC boring and milling tools according to claim 1, characterized in that, The obtained risk index satisfies the following expression: ； In the formula, This indicates the risk index at the current moment; This represents the historical average at the current moment; Represents the characteristic force at the current moment; Indicates the historical deviation at the current moment; Indicates the instantaneous deviation at the current moment; Represents the maximum value function; It represents an extremely small positive number, and the denominator is guaranteed to be non-zero.

8. The intelligent monitoring method for cutting force of CNC boring and milling tools according to claim 1, characterized in that, The process involves comparing risk indices against basic statistical thresholds to determine if they are abnormal. After processing is completed, a digital archive of the processing process is generated, including: The system compares the current risk index with the basic statistical threshold to determine whether a cutting abnormality has occurred. When a complete machining task ends, whether it is completed normally or interrupted by an alarm, the system binds the entire data of this machining task with the unique identifier of the workpiece to generate a machining process digital archive. The machining process digital archive includes: a risk index sequence; related data at the moment when the risk index exceeds the basic statistical threshold; real-time principal force, real-time transverse force, and real-time longitudinal force data within 2 seconds before and after the alarm moment; and machine tool coordinates and spindle speed at the moment of the alarm.

9. A CNC boring and milling tool cutting force intelligent monitoring system, characterized in that, include: The processor and memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a method for intelligent monitoring of cutting force of CNC boring and milling tools according to any one of claims 1-8.

Images