Intelligent prediction maintenance system for hardware tool production equipment

Abstract

The invention relates to the technical field of equipment dust hazard maintenance, in particular to a hardware tool production equipment intelligent prediction maintenance system which comprises a dust cooperative sensing unit, a dust feature library construction unit, a dust weighted decision unit and a maintenance execution unit. The dust cooperative sensing unit collects dust concentration, a vibration energy peak value and a current fluctuation rate through a three-dimensional gridding sensor group, generates a dust-vibration-current baseline comparison table through self-calibration, associates a mapping relation with the dust cumulative gradient and dust feature library construction unit, and distinguishes an equipment operation stage optimization baseline. The dust weighted decision-making unit divides three levels of dust hazard levels, dynamically adjusts a threshold value and calculates a comprehensive influence index in combination with an equipment type weight, and the maintenance execution unit receives a grading instruction and outputs a maintenance scheme containing process compensation parameters, so that the problems of insufficient correlation of traditional maintenance parameters and one-sided risk assessment are solved; the fault early warning accuracy and the maintenance pertinence are improved, and the production continuity is guaranteed.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of equipment dust hazard maintenance, in particular to a hardware tool production equipment intelligent predictive maintenance system. BACKGROUND

[0002] Equipment dust hazard maintenance is an important technology. In the hardware tool production process, the equipment is in a working condition of dust concentration and high-frequency vibration for a long time. Dust intrusion and mechanical wear can easily cause equipment failure. Precise predictive maintenance is a core means to ensure production continuity and reduce downtime losses. This technology can improve production efficiency and prolong equipment life by monitoring the equipment operating state in real time and identifying potential failure risks in advance. Traditional maintenance methods that rely on regular maintenance or single parameter monitoring have been unable to meet the needs of failure warning under complex working conditions.

[0003] However, the traditional hardware tool production equipment maintenance technology has the core problems of failure warning lag and one-sided risk assessment. Existing solutions only monitor a single parameter through a fixed threshold, without considering the correlation between dust concentration and equipment operating parameters. When dust accumulation causes bearing wear, the abnormal change of vibration signals is often ignored, and maintenance is triggered only when the equipment fails obviously, causing downtime losses. At the same time, without considering the dynamic changes of dust concentration and the differences in equipment types for risk assessment, the same warning standard is used for punch presses and welding equipment in high-dust environments, leading to risk misjudgment or excessive maintenance. These problems make equipment failures occur frequently, increasing maintenance costs and affecting production progress due to sudden downtime, which cannot meet the needs of hardware tool production for equipment maintenance precision and foresight. In order to solve this technical problem, we provide a hardware tool production equipment intelligent predictive maintenance system. SUMMARY

[0004] The purpose of the present application is to provide a hardware tool production equipment intelligent predictive maintenance system to solve the problems raised in the background art.

[0005] 1. Since traditional maintenance does not correlate dust and equipment parameters, single parameter monitoring leads to failure warning lag, therefore, the present case establishes a dust-vibration-current baseline comparison table through a dust coordination sensing unit, combined with dust accumulation gradient analysis. It can identify dust-induced failure risks in advance and avoid downtime losses.

[0006] 2. Since traditional maintenance does not have a graded risk assessment, equipment type differences are ignored, leading to misjudgment, therefore, the present case divides the dust hazard level into three levels through a dust weighting decision unit, dynamically adjusts the threshold and calculates the comprehensive influence index. It can accurately match the equipment type and risk level, and improve the pertinence of maintenance.

[0007] To achieve the above objectives, an intelligent predictive maintenance system for hardware tool production equipment is provided, including a dust collaborative sensing unit. A three-dimensional grid-distributed sensor group extracts real-time dust concentration data, vibration energy peak value, and current fluctuation rate into a dust feature library construction unit. The system is characterized by also including: During the trouble-free operation phase of the equipment, a correlation mapping relationship between dust concentration and vibration and current fluctuations is established to generate a dust-vibration-current baseline comparison table. The instantaneous rate of change of dust concentration is calculated and marked as the dust accumulation gradient. The baseline comparison table and dust accumulation gradient are output to the dust weighted decision unit. The dust weighted decision unit receives the baseline comparison table and dust accumulation gradient from the dust feature library construction unit and performs multi-dimensional analysis: The system divides dust hazard levels into three levels based on real-time dust concentration data and dust accumulation gradients. The threshold is dynamically adjusted according to the dust hazard level. The deviation between the current vibration energy peak and the baseline value is compared, and the dust comprehensive impact index is calculated in combination with the current fluctuation rate. When the dust comprehensive impact index exceeds the dynamically adjusted threshold, the equipment is determined to be in a high-risk dust state and a graded control instruction is output to the maintenance execution unit.

[0008] Compared with the prior art, the present invention has the following beneficial effects: 1. The dust collaborative sensing unit uses a three-dimensional grid sensor group to accurately collect data. Combined with self-calibration in the trouble-free operation stage and multi-stage data accumulation, the constructed dust-vibration-current baseline comparison table can dynamically reflect the correlation between the three under different working conditions. When the dust concentration changes, the dust accumulation gradient generated by the sliding window mechanism can capture the concentration change trend in real time, avoiding the omission of potential faults by traditional single parameter monitoring, and identifying hidden dangers such as bearing wear caused by dust intrusion in advance, providing accurate data support for fault warning.

[0009] 2. The three-level dust hazard level divided by the dust weighted decision unit can dynamically adjust the judgment threshold based on the real-time concentration and cumulative gradient. For different types of equipment such as stamping machines and welding equipment, a type weight factor is introduced to calculate the comprehensive dust impact index, making the risk assessment more in line with the characteristics of the equipment, avoiding the waste of resources caused by excessive maintenance and preventing sudden failures caused by misjudgment of risks.

[0010] 3. Hierarchical control instructions contain process compensation parameters adapted to the equipment type and are executed after virtual verification by the digital twin model, ensuring the effectiveness and safety of maintenance measures, reducing production interruptions caused by improper maintenance, extending equipment service life, and improving the continuity and stability of hardware tool production. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is an overall block diagram of the present invention.

[0012] The meanings of the various reference numerals in the figures are as follows: 1, dust collaborative sensing unit; 2, dust feature library construction unit; 3, dust weighted decision unit; 4, maintenance execution unit. DETAILED DESCRIPTION

[0013] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0014] The present application provides a hardware tool production equipment intelligent predictive maintenance system, please refer to Figure 1 As shown in the figure, it comprises a dust collaborative sensing unit 1, which extracts real-time dust concentration data, vibration energy peak value and current fluctuation rate to the dust feature library construction unit 2 through a three-dimensional grid distribution of sensor groups.

[0015] In order to realize accurate monitoring of dust and running parameters of hardware tool production equipment, the dust collaborative sensing unit 1 optimizes sensor deployment and self-calibration process to construct baseline data and dynamic gradient index that fit the characteristics of the equipment, and the specific implementation is as follows: The three-dimensional grid distribution of sensor groups in the dust collaborative sensing unit 1 is deployed by using spatial topology optimization algorithm, and the sensor groups are arranged differently according to the structural characteristics and dust accumulation law of different equipment: A hemispherical monitoring network is formed at the hydraulic station of the punch, cross-shaped monitoring lines with equal intervals are set at the welding station, axial gradient distribution sensors are installed at the main shaft of the cutting machine, 8 sensors are arranged within a range of 3 meters around the hydraulic station to form a hemispherical monitoring network, with the core components of the hydraulic station as the center of the sphere, the sensors are evenly distributed in the upper half of the spherical surface, 1-2 meters away from the center of the sphere, to ensure that the hydraulic pump, oil pipe interface and other parts prone to dust leakage and high vibration are covered, and the dust concentration and vibration transmission characteristics in different directions can be captured at the same time. During the welding process, dust spreads linearly along the airflow, cross-shaped monitoring lines with equal intervals are set along the X axis and Y axis of the workbench, one sensor is installed every 50 centimeters on each line, and the sensor density is doubled at the intersection, focusing on monitoring the dust concentration and current fluctuation around the welding torch and near the smoke collector. Welding current is easily affected by dust short circuit, dust spreads along the axial direction when the main shaft rotates at high speed, and vibration energy is transmitted along the shaft, so one sensor is installed every 10 centimeters on the main shaft shell from the front end to the rear end to form an axial gradient distribution, accurately capturing the variation law of dust concentration with the length of the main shaft and the vibration energy attenuation characteristics at different positions, so that the sensors are "close" to the dust source and parameter sensitive area, ensuring that the collected data can truly reflect the running state of the key parts of the equipment; The sensor group performs self-calibration in the first fault-free operation stage of the equipment, synchronously collects the original data of dust concentration, vibration energy and current fluctuation rate, removes abnormal values caused by the impact of equipment start and stop, arranges the effective data in the format of "dust concentration interval-corresponding vibration energy range-corresponding current fluctuation range", and stores them into the dust-vibration-current baseline comparison table. For example, when the dust concentration is 5-10 mg / m3, the vibration energy peak of the hydraulic station of the punch press should be 50-80 energy units, and the current fluctuation rate should be 2%-5%. These data are used as the baseline for subsequent judgment of equipment abnormalities, and the sliding window mechanism is used to calculate the dust concentration change rate. The sliding window length is set to 1 minute, and the sliding interval is 30 seconds, that is, 2 gradient values are generated every minute, which can capture short-term changes and avoid data redundancy. For the dust concentration data in each window, the concentration difference between the end and the beginning of the window is calculated, such as 8 mg / m3 at the beginning and 12 mg / m3 at the end, with a difference of 4 mg / m3. The concentration change is divided by the window length to obtain the concentration change rate per minute, that is, the dust accumulation gradient, such as 4 mg / m3 ÷ 1 minute = 4 mg / (m3·min).

[0016] If the gradient value is positive and continuously increasing, it indicates that the dust is accelerating accumulation, for example, the dust concentration in a window of a welding station increases from 15 mg / m3 to 21 mg / m3, with a gradient value of 6 mg / (m3·min). The system will mark that the dust accumulation speed in this area is fast and needs to be paid attention to, and generate a dynamic dust accumulation gradient, avoiding the one-sidedness of traditional single sensor monitoring.

[0017] characterized in that it further comprises: In the fault-free operation stage of the equipment, the correlation mapping relationship between dust concentration and vibration and current fluctuation is established, the dust-vibration-current baseline comparison table is generated, and the instantaneous change rate of dust concentration is calculated, which is marked as dust accumulation gradient. The baseline comparison table and the dust accumulation gradient are output to the dust weighted decision unit 3; To accurately construct dust-vibration-current correlation baseline adapted to different working conditions, the fault-free operation stage generates a three-dimensional baseline comparison table with working condition adaptability through phased data collection and multi-domain analysis. The specific implementation is as follows: The fault-free operation stage is divided into three stages: running-in period, stable period and aging verification period. The running-in period is for the state of the equipment just put into operation, the parts have not completely adapted, and the focus is to collect the relevant data under light load conditions. The running-in period collects the dust concentration and vibration energy correlation data of the equipment under typical light load conditions, establishes the basic operation characteristics, and defines the typical light load conditions according to the equipment type. For example, the punch machine sets the punch frequency to 30% of the rated value, such as 10 times per minute normally, and the light load is 3 times. The welding equipment current is adjusted to 50% of the rated value, such as 200 amperes normally, and the light load is 100 amperes. The cutting machine spindle speed is reduced to 40% of the rated value. Under the light load condition, the sensor group continuously records the dust concentration and vibration energy peak value for 48 hours.For example, the hydraulic station of the punch press is under light load, data is collected once an hour, and when the dust concentration is 3 mg / m3, the vibration energy peak value is stable at 40 energy units; when the concentration rises to 8 mg / m3, the vibration energy peak value increases to 60 energy units. The data is sorted according to the corresponding relationship of “dust concentration-vibration energy” to form the basic characteristic curve under the light load condition, which is used as a reference for the initial stable operation of the equipment. As shown in the curve, the vibration energy increases linearly with the increase of dust concentration under light load, with a slope of 5 energy units / mg / m3. The stable period is extended to the conventional production load range to obtain the dynamic mapping relationship between dust concentration and current fluctuation rate. The stable period is focused on the stage when the equipment enters the normal production state, and is extended to the conventional production load range to supplement the current fluctuation rate data. The load of the equipment is gradually adjusted to 60%-100% of the rated value. For example, the punch press is operated at 60%, 80%, and 100% of the rated frequency, and each load level is continuously operated for 24 hours to ensure that the data covers the typical working conditions of daily production. At each load level, the changes of dust concentration and current fluctuation rate are recorded simultaneously. For example, when the welding equipment is under 80% load, the dust concentration is 10 mg / m3 and the current fluctuation rate is 3% when the dust concentration is 10 mg / m3. When the concentration rises to 15 mg / m3, the fluctuation rate increases to 6%. At the same time, whether the vibration energy changes synchronously is observed. For example, the vibration energy increases from 70 to 90 energy units. The corresponding relationship among “dust concentration-vibration energy-current fluctuation rate” is included in the baseline data to form a dynamic mapping table containing different loads. As shown in the table, under 80% load, the dust concentration of 10-15 mg / m3 corresponds to the vibration energy of 70-90 energy units and the current fluctuation rate of 3%-6%. The aging verification period records the nonlinear response characteristics of vibration energy by simulating the sudden change of dust concentration under extreme working conditions. For example, by closing part of the dust removal equipment, the dust concentration is suddenly increased from 20 mg / m3 to 50 mg / m3 within 10 minutes, and the load of the equipment is adjusted to 110% of the rated value (simulating short-term overload). The parameter changes are observed, and the abnormal changes of vibration energy are monitored. For example, when the dust concentration suddenly increases, the vibration energy of the spindle of the cutting machine does not increase linearly, but suddenly appears high-frequency pulse, such as from 100 energy units to 180 energy units, and then decreases. The jump of the current fluctuation rate is recorded. The critical point of the nonlinear change of the parameters under extreme conditions is recorded, such as when the dust concentration is 40 mg / m3, the vibration energy starts to pulse abnormally, which is used as the implicit threshold of high-risk early warning. Recording the response characteristics can make the baseline comparison table cover the full range of “safety-risk” to avoid the failure of early warning in extreme cases. The three-stage data is analyzed by time domain-frequency domain joint analysis, that is, the change law and frequency distribution characteristics of the data over time are analyzed at the same time. Time domain analysis is used to extract the change trend of the parameters, such as the increase of 2 mg / m3 per hour of dust concentration in the stable period, and frequency domain analysis is used to identify characteristic frequencies, such as the normal vibration frequency of 50 Hz of the hydraulic station of the punch machine, and the high-frequency component of 100 Hz appears after the dust intrusion. The effective data after analysis is arranged according to the structure of "operation stage-load level-dust concentration interval-vibration energy range-current fluctuation rate range", and a three-dimensional baseline comparison table with working condition adaptability is generated. The table stores the correlation threshold of dust concentration, vibration and current fluctuation under different operating conditions, which provides a precise benchmark for the subsequent dust weighted decision unit 3 to judge whether the equipment is abnormal.

[0018] The dust weighted decision unit 3 receives the baseline comparison table of the dust feature library construction unit 2 and the dust accumulation gradient, and performs multi-dimensional analysis: According to the real-time dust concentration data and the dust accumulation gradient, the dust hazard level is divided into three levels, and the threshold is dynamically adjusted according to the dust hazard level. The deviation of the current vibration energy peak value from the baseline value is compared, and the dust comprehensive influence index is calculated combined with the current fluctuation rate. When the dust comprehensive influence index exceeds the dynamically adjusted threshold, it is determined that the equipment is in a high-risk state of dust, and the grading control instruction is output to the maintenance execution unit 4.

[0019] In order to accurately extract the vibration energy abnormal characteristics of the hardware tool production equipment caused by dust intrusion, the improved time-frequency domain hybrid analysis method is used to extract the vibration energy peak value, and the key characteristics are separated and the abnormal pulse is tracked through multiple steps of signal processing. The specific implementation is as follows: The original vibration signal is decomposed by wavelet packet, and the metal friction characteristic component in the high frequency band is separated. The decomposition process divides the signal into multiple frequency bands from low to high frequency, such as 1-100 Hz for low frequency, 100-500 Hz for medium frequency, and 500-2000 Hz for high frequency. The signal in the high frequency band is separated, which corresponds to the vibration caused by metal friction, such as the friction between bearing balls and tracks caused by dust wear. It is called metal friction characteristic component. For example, the original vibration signal of the hydraulic station of the punch machine is decomposed, and the signal in the high frequency band of 500-800 Hz is extracted separately. This signal can reflect the friction state of the hydraulic pump bearing. The logic of this step is: The equipment wear caused by dust invasion is mainly manifested as high-frequency vibration. By decomposing and separating the high-frequency signal, the low-frequency vibration interference generated during normal operation of the equipment, such as the basic vibration of motor rotation, can be excluded, and the focus is on the wear-related characteristic information. Based on the separated high-frequency metal friction characteristic component, the root mean square value of the signal in the preset time window is calculated as the vibration energy benchmark, and the energy difference of adjacent windows is calculated to calculate the instantaneous gradient. The high-frequency signal is divided into a time window every 0.1 seconds, that is, each window contains the vibration data within 0.1 seconds, which ensures that the energy change in a short time can be captured. For the high-frequency signal in each time window, the root mean square value is calculated, which is a statistical value reflecting the energy strength of the signal. The value is taken as the vibration energy benchmark of the current window. For example, if the root mean square value of a certain window is 80, it represents the vibration energy level within 0.1 seconds. The energy benchmark of the current window is subtracted from the energy benchmark of the previous window to obtain the energy difference of adjacent windows. For example, the current window is 80, the previous window is 70, and the difference is 10. The difference is the instantaneous gradient, which is used to reflect the speed of energy change. For example, in the high-frequency signal of the main shaft of the cutting machine, the energy benchmarks of the last three windows are 90, 95 and 110 respectively, and the corresponding instantaneous gradients are 5 and 15, indicating that the vibration energy is rising rapidly, which may be related to the intensification of dust wear. When the dust accumulation gradient exceeds the preset warning value, the high-frequency energy tracking algorithm is activated to capture the bearing micro-vibration pulse waveform caused by dust invasion. The sampling frequency of the high-frequency signal is increased from the original 1000 times per second to 5000 times per second to capture more subtle vibration changes. The algorithm filters out the sharp pulses that suddenly appear in the high-frequency signal through the waveform recognition module, such as the energy rising from 100 to 200 in an instant and then falling rapidly. Such pulses are the micro-vibration characteristics of the bearing caused by dust invasion. Dust particles get stuck between the ball and the track, causing instantaneous impact vibration. The captured pulse waveform is recorded for its occurrence time, peak energy and duration, such as a peak value of 220 appearing at 10:05:30 and a duration of 0.001 seconds, which serves as a basis for judging the degree of bearing wear. For example, when the dust accumulation gradient of the welding equipment reaches 6 mg / (m3·min), which exceeds the warning value of 5, the tracking algorithm is activated, and three consecutive sharp pulses are found in the high-frequency signal, indicating that the bearing has been affected by dust and has abnormal wear, which needs to be warned in time. This provides a reliable vibration parameter basis for subsequent risk level classification.

[0020] To accurately calculate the current fluctuation rate and eliminate external interference, the current fluctuation rate calculation process incorporates a hardware compensation mechanism to generate a reliable deviation index through anti-interference processing and dynamic comparison. The specific implementation is as follows: When the dust sensor detects a concentration exceeding the preset dust concentration threshold, the current ripple detection module starts the anti-interference mode, filters out the periodic harmonic components caused by the fluctuations of the workshop power grid, and the module identifies the periodic harmonic components generated by the fluctuations of the workshop power grid through spectral analysis, such as the multiple frequencies of the power grid frequency of 50 Hz, such as 100 Hz, 150 Hz interference signals. These components will cause irregular fluctuations in the current detection value, which is unrelated to the true current change caused by dust. Enable band-stop filter to filter out the identified harmonic components, and retain the fundamental current signal generated by the normal operation of the device, such as the working current fundamental frequency of 60 Hz of the welding device, to ensure that the subsequent calculation is based on pure current data. For example, the dust concentration of a certain welding station reaches 25 mg / m³, exceeding the preset threshold of 20, and the current ripple detection module detects 100 Hz harmonic interference introduced by the power grid. The band-stop filter filters out the signal at this frequency, so that the current data only reflects the true fluctuations of the welding process. On the basis of anti-interference processing, the moving average algorithm is used to calculate the standard deviation of the current effective value in the recent preset time period, and the calculation window length is determined according to the device type, such as 5 minutes for a punch press, and 2 minutes for a welding device, because the current of the hydraulic system changes slowly. The welding current changes more frequently with the welding point. In the preset time period, the current effective value of the device is collected at a frequency of 2 times per second, which is the equivalent DC value of the current, used to represent the average size of the current. For example, 240 current effective value data are collected in 2 minutes for a welding device. The collected current effective value is processed by the moving average algorithm. First, the average value in this time period is calculated, then the deviation of each data from the average value is calculated, and finally the standard deviation is obtained, which reflects the dispersion degree of the current data. The larger the value, the more intense the fluctuations. For example, the current effective value of the cutting machine in the 5-minute preset time period is 10 amperes, and the standard deviation is 0.5 amperes, indicating that the fluctuation amplitude of the current around the average value in this period is small. Compared with the historical data in the same dust concentration interval in the baseline comparison table, a fluctuation rate deviation index in percentage form is generated. The current dust concentration interval is retrieved from the baseline comparison table, such as 20-30 mg / m³. The historical current effective value standard deviation is, for example, 0.4 amperes. The difference between the current standard deviation and the historical standard deviation is calculated, and then the difference is converted to percentage form, i.e. deviation percentage = (current standard deviation - historical standard deviation) ÷ historical standard deviation x 100%), to obtain the fluctuation rate deviation index. For example, the current dust concentration is 25 mg / m³ (belonging to the 20-30 interval), the current effective value standard deviation is 0.6 amperes, and the historical standard deviation in the baseline comparison table is 0.4 amperes. The deviation percentage is (0.6-0.4) ÷ 0.4 x 100% = 50%, i.e. the index of "fluctuation rate deviation 50%" is generated, which provides accurate current parameter basis for subsequent dust comprehensive influence index calculation.

[0021] To accurately classify the dust hazard level of hardware tool production equipment, the time-space correlation analysis technique is adopted to realize dynamic determination of three-level dust hazard level, combined with the time-space characteristics of equipment operation, i.e. the trend of dust concentration changing with time and the spatial correlation of different parameters. The specific implementation is as follows: Low-risk level determination: when the real-time dust concentration data is lower than the upper limit of the safety threshold marked in the dust-vibration-current baseline table, and the change amplitude of the dust accumulation gradient in the continuous three monitoring periods does not exceed the historical average fluctuation range recorded in the baseline table, the current threshold is maintained unchanged; The real-time collected dust concentration data, such as the detection of 8 mg / m3 in the hydraulic station of the punch press, is lower than the upper limit of the safety threshold marked in the dust-vibration-current baseline table. For example, if the upper limit of the safety threshold of this equipment is 15 mg / m3, it indicates that the current dust concentration does not pose a significant threat to the equipment. In the continuous three monitoring periods, each period is 5 minutes, the dust accumulation gradient, i.e. the fluctuation amplitude of the concentration change per minute, does not exceed the historical average fluctuation range recorded in the baseline table, such as the historical average fluctuation range of ±1 mg / (m3·min). The gradients of the current three periods are 1.2, 1.0, and 1.1 mg / (m3·min), respectively, all within the range. When the above conditions are met at the same time, it is determined as low-risk level, and the determination thresholds of vibration energy peak value and current fluctuation rate are maintained unchanged, such as maintaining the vibration threshold at 80 energy units and the current fluctuation rate threshold at 5%. The logic is that in the low-risk state, the equipment parameters are stable, and there is no need to adjust the threshold to accurately monitor, avoiding excessive intervention affecting the production.

[0022] Medium-risk level determination: when the real-time dust concentration data exceeds the upper limit of the safety threshold but does not reach the emergency shutdown threshold in the baseline table, or the dust accumulation gradient appears a single-period mutation amplitude exceeding the historical average fluctuation range, the threshold tightening mechanism is started to lower the determination thresholds of vibration energy peak value and current fluctuation rate; Case one: the real-time dust concentration exceeds the upper limit of the safety threshold, such as reaching 18 mg / m3, but does not reach the emergency shutdown threshold in the baseline table, such as the emergency shutdown threshold being 30 mg / m3, indicating that the dust concentration is high but not at the dangerous critical value; Case two: the mutation amplitude of the dust accumulation gradient in a single monitoring period exceeds the historical average fluctuation range, such as the historical average fluctuation being ±1 mg / (m3·min), and the current period gradient increasing from 1.0 to 2.5 mg / (m3·min), with a mutation amplitude of 1.5 exceeding the range, indicating that the dust concentration growth rate is abnormal and may exceed the standard in the short term; When any of the above conditions are met, the vibration energy peak and current fluctuation rate determination threshold is automatically lowered. The lowering logic is: in the medium risk state, the sensitivity of the equipment to dust increases, the determination standard needs to be lowered to capture abnormal changes in vibration and current earlier, for example, when the welding station dust concentration reaches 22 mg / m3, which exceeds the safety threshold of 20 but does not reach the shutdown threshold of 35, it is determined to be a medium risk, the vibration threshold is lowered from 90 to 80, and the current fluctuation rate threshold is lowered from 6% to 5%, ensuring early identification of potential risks.

[0023] High risk level determination: when the real-time dust concentration data exceeds the emergency shutdown threshold for a continuous preset period of time, and the following conditions are met simultaneously, a dust high-risk state is triggered: (a) The current vibration energy peak exceeds the upper limit of the vibration threshold in the baseline control table corresponding to the dust concentration interval, such as the upper limit of the vibration threshold in the 35 mg / m3 interval being 120 energy units, and the actual detection being 130 energy units, indicating that the dust has caused abnormal wear of the equipment components; (b) The deviation percentage of current fluctuation rate from baseline value exceeds the preset deviation percentage of the recent maximum fluctuation value recorded by the dust feature library construction unit 2, such as the recent maximum fluctuation value being 10% and the preset deviation percentage being 80%, i.e. the current deviation needs to be ≥8%, and the actual detection being 9%, indicating that the current fluctuation is close to the historical highest risk level; (c) The dust accumulation gradient shows a continuous upward trend and the dust removal equipment has not been triggered, such as the gradient being 2.5, 3.0, 3.5 mg / (m3·min) for 3 consecutive periods, and the system detecting that the dust removal equipment has not started, such as the dust removal fan current being 0, indicating that it is not running or running but not effective, such as the fan running but the dust concentration still rising, indicating that external intervention has failed, and the risk continues to escalate, when the above basic conditions and three simultaneous conditions are met, the equipment is determined to be in a dust high-risk state, and the highest level of early warning signal is immediately output, such as the cutting machine spindle dust concentration reaching 40 mg / m3 for 10 minutes, exceeding the emergency shutdown threshold of 30, the vibration energy being 140, exceeding the threshold of 120, the current fluctuation rate deviation being 9%, exceeding the recent maximum fluctuation of 10% by 80%, and the dust removal equipment not starting, the gradient continuously rising, it is determined to be a high risk, triggering an emergency shutdown warning, making the risk assessment more in line with the actual operation state of the equipment, and realizing a step-by-step warning from safe to dangerous, providing accurate basis for maintenance decision-making.

[0024] To dynamically adapt the threshold to the health state change of the equipment, the dynamic adjustment of the threshold adopts an equipment health degree feedback closed loop mechanism, through the cooperative operation of initial setting, positive compensation and negative adjustment, to ensure that the threshold always fits the actual running ability of the equipment, the specific implementation is as follows: The initial threshold is set based on the statistical average value of the baseline table, the statistical average value of the parameters of the equipment under the normal load in the stable period is extracted from the baseline table, such as the safe average value of the dust concentration of the hydraulic station of the punch press 10 mg / m3, the average value of the vibration energy 70 energy units, the average value of the current fluctuation rate 3%, these average values are set as the initial threshold value in a certain proportion, such as the initial threshold value of the dust concentration 12 mg / m3, the initial threshold value of the vibration energy 84 energy units, the initial threshold value of the current fluctuation rate 3.6%, the setting logic of the initial threshold value is: based on the average state of the normal operation of the equipment, a certain safety margin is reserved, which can avoid misjudgment and timely alarm when the parameters approach abnormality. For example, the average value of the current fluctuation rate in the baseline table of the welding equipment is 4%, the initial threshold value is set to 4.8%, to ensure that normal fluctuation will not trigger the early warning, when the cumulative running time of the equipment reaches the maintenance period, the threshold value is positively compensated, according to the maintenance record of the equipment, if the parameter recovery effect is good after historical maintenance, such as the vibration energy decreases by 20%, the compensation amplitude is set to 5%-10% of the initial threshold value, such as the initial threshold value of the vibration energy is 84, after 10% compensation, it becomes 92.4, the compensated threshold value is used as the new judgment standard, such as the dust concentration threshold value is increased from 12 mg / m3 to 13.2 mg / m3, the logic of positive compensation is: after maintenance, the health degree of the equipment is improved, the running stability is enhanced, the threshold value can be appropriately relaxed to avoid excessive early warning due to the use of the original standard after the performance of the equipment is restored; When the dust accumulation gradient enters the high risk level, the threshold negative compensation algorithm is activated to adjust the threshold value, and the single adjustment amplitude is limited according to the baseline value, the single adjustment amplitude is set to 3%-5% of the baseline value according to the severity of the high risk level, and the threshold value after single adjustment cannot be lower than 80% of the baseline value, to avoid misjudgment caused by excessive reduction, the threshold values of the vibration energy peak value and the current fluctuation rate are reduced, such as the vibration threshold value is reduced from 84 to 80.5, the current fluctuation rate threshold value is reduced from 4.8% to 4.6%, the logic of negative compensation is: when the dust accumulation gradient enters the high risk level, it indicates that the equipment may have been eroded by dust, the health degree is decreased, the sensitivity to abnormal parameters needs to be improved, by reducing the threshold value, the subtle abnormalities of vibration and current can be captured earlier, such as the vibration caused by slight wear of the bearing. For example, the dust accumulation gradient of the spindle of the cutting machine reaches 6 mg / (m3·min) (high risk), the vibration threshold value is reduced from 90 to 87, to ensure timely identification of early wear, the threshold value is relaxed for new equipment or equipment after maintenance, the threshold value is tightened in aging or high risk state, which ensures the timeliness of early warning and avoids unnecessary production interruption, making the risk judgment more in line with the actual health level of the equipment.

[0025] To accurately quantify the comprehensive risk of different types of equipment affected by dust, the calculation of the dust comprehensive influence index introduces a device type weight factor, and a standardized risk index is generated by combining data fusion strategies. The specific implementation is as follows: According to the structural characteristics and dust sensitive parameters of different equipment, different weight factors are assigned to each type of equipment, highlighting the core influence parameters. The vibration energy deviation of the punch machine is the dominant factor, and the wear of its core components (hydraulic pump, bearing) is mainly reflected through vibration energy. Therefore, the vibration energy deviation is the dominant factor, and the weight of the difference between the current value and the baseline value in the comprehensive index is 60%, the weight of the dust concentration deviation is 30%, and the weight of the current fluctuation rate deviation is 10%. For example, if the vibration energy deviation of the punch machine is 20 energy units, the dust concentration deviation is 5 mg / m3, and the current fluctuation rate deviation is 2%, each parameter will participate in the calculation according to the proportions of 60%, 30%, and 10%. The welding equipment emphasizes the contribution of current fluctuation, and the welding quality is highly related to the stability of the current. Dust can easily cause electrode short circuit and induce current fluctuation, so the contribution of current fluctuation is emphasized. The weight of the current fluctuation rate deviation in the comprehensive index is 60%, the weight of the dust concentration deviation is 25%, and the weight of the vibration energy deviation is 15%. For example, if the current fluctuation rate deviation of the welding equipment is 4%, the dust concentration deviation is 8 mg / m3, and the vibration energy deviation is 10 energy units, the weighted calculation is performed according to the above proportions. The cutting machine introduces the main shaft working condition attenuation coefficient. When the main shaft rotates at high speed, dust intrusion will aggravate the wear of the main shaft, and the wear degree accumulates with running time. Therefore, the main shaft working condition attenuation coefficient is introduced to reflect the degree of performance decline caused by wear. This coefficient is multiplied by the dust concentration deviation and the vibration energy deviation to participate in the calculation, with a weight distribution of 40% for the dust concentration deviation, 40% for the vibration energy deviation, and 20% for the current fluctuation rate deviation. At the same time, the main shaft working condition attenuation coefficient is dynamically adjusted according to the cumulative running time of the main shaft. The longer the running time, the larger the coefficient. For example, if the running time is 100 hours, the coefficient is 1.0, and if the running time is 500 hours, the coefficient is 1.2. This allows the index calculation to focus on the core parameters of the equipment that are most susceptible to dust, ensuring that the risk assessment of different types of equipment is more in line with their structural characteristics. The calculation process asymmetrically fuses real-time dust data with historical status over a rolling time window, with recent data holding the dominant weight and historical data holding the secondary weight. The 10-minute data within the window is divided into two segments: the recent data (the last 3 minutes) and the previous 7 minutes. Recent data holds the dominant weight (70%), while historical data holds the secondary weight (30%). For example, if the recent dust concentration deviation within the window is 6 mg / m³ and the historical average deviation is 4 mg / m³, the fused dust concentration deviation is calculated as 6 × 70% + 4 × 30%. The window is rolled forward every 5 minutes, discarding the oldest 5 minutes of data and including the latest 5 minutes of data to ensure that the window always contains the last 10 minutes of information. For example, the window from 10:00-10:10 rolls over to 10:05-10:15 at 10:05. Recent data from 10:05-10:10 is retained, while real-time data from 10:10-10:15 is added. The logic behind asymmetric fusion is that recent data better reflects the current state of the equipment, and increased weighting allows the index to quickly respond to the latest risk changes. Historical data, used as a reference, can help prevent index errors caused by short-term fluctuations. The weighted deviations of each parameter are aggregated and normalized to produce a standardized dust risk index. The aggregate value is the sum of the weighted results of each parameter, and the index threshold range is set: 0-30 for low risk, 30-70 for medium risk, and 70-100 for high risk. For example, the aggregate value for a stamping machine is 65 after normalization, resulting in a comprehensive impact index of 65, which is considered medium risk. This provides accurate risk quantification for subsequent maintenance execution unit 4. A higher index indicates a more severe dust impact on the equipment, requiring priority maintenance.

[0026] To accurately verify the deviation between the vibration energy peak and baseline values ​​and ensure the accuracy of dust risk assessment, the vibration energy deviation comparison adopts a multi-band coupling verification method. By separating the different frequency bands of the vibration signal and monitoring key impact characteristics, it can achieve accurate identification of equipment abnormal conditions. The specific implementation method is as follows: First, the vibration signal collected by the sensor is divided into frequency bands. According to the correlation characteristics of signal frequency and equipment fault type, three core frequency bands are separated, and the vibration signal is decomposed into a base frequency band, a friction characteristic band and a dust impact band. The base frequency band corresponds to the basic vibration of the equipment during normal operation, such as the inherent vibration of motor rotation and hydraulic pump operation, with a frequency range of 20-100 Hz. The energy of this band is stable, and the deviation is usually caused by changes in equipment load, with low correlation with dust. The friction characteristic band corresponds to the vibration generated by the friction of equipment components (such as bearings and gears), with a frequency range of 100-500 Hz. Dust intrusion will exacerbate friction, causing the energy of this band to abnormally increase, which is a key frequency band for judging mechanical wear. The dust impact band corresponds to the instantaneous vibration generated by the high-speed collision of dust particles in the gap between equipment, with a frequency range of 500-2000 Hz. The signals in this band show pulse characteristics, which are typical signs of direct dust intrusion. For example, after separating the vibration signal of a punch machine hydraulic station, the base frequency band (20-100 Hz) reflects the vibration of the normal operation of the hydraulic pump, the friction characteristic band (100-500 Hz) reflects the friction state of the piston and cylinder, and the dust impact band (500-2000 Hz) reflects the impact of dust particles on the seal. Calculate the deviation rate of each frequency band energy from the baseline value. For the vibration signal of each frequency band, calculate the energy value (characterized by the sum of the squares of the signal amplitude) within a predetermined time window (such as 1 minute). From the dust-vibration-current baseline table, obtain the energy baseline value of each frequency band under the current equipment type and current dust concentration interval. For example, when the dust concentration of the punch machine is 10-20 mg / m3, the base frequency band baseline value is 50 energy units, the friction characteristic band is 80 energy units, and the dust impact band is 30 energy units. The deviation rate of each frequency band is (current energy value - baseline value) ÷ baseline value x 100%. For example, the current energy value of the friction characteristic band is 100 energy units, and the baseline value is 80, so the deviation rate is (100-80) ÷ 80 x 100% = 25%. Through the calculation of multi-band deviation rate, different types of vibration anomalies can be distinguished: a large deviation in the base frequency band may be load fluctuation, a large deviation in the friction characteristic band may be mechanical wear, and a large deviation in the dust impact band may be direct dust intrusion. When the dust concentration enters the high-risk level, the transient impact times of the dust impact section are monitored. If there are n times of pulse waveforms with amplitudes exceeding the baseline value safety range within a preset time, it is directly determined that the dust is in a high-risk state. The preset monitoring time window and impact times threshold n are set according to the type of the equipment (for example, n=5 times for a punch machine and n=3 times for a welding device). The baseline value safety range is also specified (for example, the baseline value of the dust impact section is 30 energy units, and the safety range is 30±5 energy units, that is, a pulse exceeding 35 energy units is considered abnormal). Within the monitoring time window, the signals of the dust impact section are tracked in real time, and transient pulse waveforms with amplitudes exceeding the safety range are identified. The time and amplitude of each impact are recorded. If the number of abnormal pulses within 5 minutes reaches or exceeds n times, it is directly determined that the equipment is in a high-risk dust state, and there is no need to wait for verification of other parameters. The core of this determination logic is that the abnormal pulse of the dust impact section is the "immediate evidence" of the direct damage of dust to the equipment. The frequent occurrence of such pulses under the high-risk level indicates that the equipment has been severely eroded by dust, and immediate warning needs to be triggered to avoid failure. For example, when the dust concentration of a cutting machine enters the high-risk level (40 mg / m3), 7 pulses with amplitudes exceeding 40 energy units are detected in the dust impact section within 5 minutes (baseline safety range 30±5), which exceeds the set n=5 times, and it is directly determined that the dust is in a high-risk state. An urgent maintenance instruction is output. Through multi-frequency coupling verification, different sources of vibration signals are analyzed comprehensively, and the rapid response to emergency situations is realized through impact monitoring under the high-risk level. The comparison result of vibration energy deviation more accurately reflects the actual harm of dust to the equipment.

[0027] To make the hierarchical control instructions accurately adapt to the type and risk state of the equipment, the instructions output by the maintenance execution unit 4 contain targeted process compensation parameters, which are executed after virtual verification. The specific implementation is as follows: For the hydraulic pressure attenuation coefficient output by the hydraulic station of the punch press, the main cylinder pressure is reduced according to the dust risk index. A mapping table of the preset dust risk index interval and the attenuation coefficient is set. The index 0-30 (low risk) corresponds to the attenuation coefficient 1.0 (no pressure reduction), 30-70 (medium risk) corresponds to 0.9 (10% reduction), and 70-100 (high risk) corresponds to 0.8 (20% reduction). For example, the dust risk index of a certain punch press is 60 (medium risk), and the matching attenuation coefficient is 0.9. The compensated pressure value is obtained by multiplying the current main cylinder pressure by the attenuation coefficient. For example, if the current main cylinder pressure is 10 MPa, it is adjusted to 9 MPa after being multiplied by 0.9. The instruction content clearly adjusts the time (such as "immediately reduce the main cylinder pressure from 10 MPa to 9 MPa, and continue until the risk index is below 30"). It ensures that the friction of the hydraulic system is reduced by reducing the pressure, and the dust invasion will exacerbate the wear. Reducing the pressure can reduce the load on the components. A protective gas flow compensation curve is generated for the welding equipment to achieve dynamic adjustment of the gas flow. The current fluctuation rate is set to increase by 2% for every 1% increase. At the same time, the upper limit of the flow is set. For example, if the current fluctuation rate of the welding equipment is 6%, which is 4% higher than the baseline value, the gas flow needs to be increased by 8. The curve takes time as the horizontal axis and flow value as the vertical axis, and marks the target flow at different time points, such as 0-5 minutes of flow increase of 5%, and 5-10 minutes of flow adjustment to 8% according to the real-time fluctuation rate. It ensures that the flow is adapted in real time with the current fluctuation. The instruction contains curve parameters, "protective gas flow increases by 2% per 1 minute until the fluctuation rate returns to within 3%", and the dust in the welding area is blown away by increasing the gas flow. Dust can cause current fluctuation, and increasing protective gas can reduce dust interference. The proportion of gas flow increase is positively correlated with current fluctuation rate. After all instructions are virtually verified through the device digital twin model, they are written into the PLC controller. The pressure adjustment value of the punch press and the flow curve of the welding equipment are input into the digital twin model, which is a virtual replica of the device and can simulate the operating state. The model simulates the state of the device after executing the instructions, such as the change in vibration energy after the pressure of the punch press is reduced and the current stability after the flow of the welding equipment is increased. It verifies whether there are any abnormalities. After virtual verification without abnormalities, the instructions are converted into codes that can be recognized by the PLC controller (Programmable Logic Controller). They are sent to the device through the industrial bus to trigger the actuator to adjust the parameters. For example, the punch press pressure adjustment instruction is verified by the twin model, confirming that the punch precision is qualified and the vibration energy is reduced by 15% at 9 MPa pressure. It is written into the PLC controller for execution. Through this process, the hierarchical control instruction contains targeted process compensation parameters (adapt to the type and risk level of the device) and is verified to ensure effectiveness through virtual verification, realizing a closed loop of "early warning-compensation-execution". It reduces the harm of dust while ensuring production continuity.

[0028] In the present invention, the dust collaborative sensing unit 1 collects dust concentration, vibration energy peak and current fluctuation rate through a three-dimensional grid sensor group, and generates a dust-vibration-current baseline comparison table and a dust accumulation gradient through self-calibration. The dust feature library construction unit 2 associates the three mapping relationships, distinguishes the equipment operation stage and optimizes the baseline. The dust weighted decision unit 3 divides the dust hazard level into three levels, dynamically adjusts the threshold and calculates the comprehensive impact index based on the equipment type weight. The maintenance execution unit 4 receives the graded instructions and outputs a maintenance plan containing process compensation parameters, which solves the problems of insufficient correlation of traditional maintenance parameters and one-sided risk assessment, improves the accuracy of fault warning and the targeted maintenance, and ensures production continuity.

[0029] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. An intelligent predictive maintenance system for hardware tool production equipment, comprising a dust collaborative sensing unit (1), which extracts real-time dust concentration data, vibration energy peak value, and current fluctuation rate to a dust feature library construction unit (2) through a three-dimensional grid-distributed sensor group, characterized in that: Also includes: During the trouble-free operation phase of the equipment, a correlation mapping relationship between dust concentration and vibration and current fluctuations is established to generate a dust-vibration-current baseline comparison table, and the instantaneous rate of change of dust concentration is calculated and marked as the dust accumulation gradient. The baseline comparison table and the dust accumulation gradient are output to the dust weighted decision unit (3); The dust weighted decision unit (3) receives the baseline comparison table and dust accumulation gradient from the dust feature library construction unit (2) and performs multi-dimensional analysis: The dust hazard level is divided into three levels according to the real-time dust concentration data and the dust accumulation gradient, and the threshold is dynamically adjusted according to the dust hazard level. The deviation between the current vibration energy peak and the baseline value is compared, and the dust comprehensive impact index is calculated in combination with the current fluctuation rate. When the dust comprehensive impact index exceeds the threshold after dynamic adjustment, it is determined that the equipment is in a high-risk dust state and a graded control instruction is output to the maintenance execution unit (4).

2. The intelligent predictive maintenance system for hardware tool production equipment according to claim 1, characterized in that: The three-dimensional grid-distributed sensor group in the dust cooperative sensing unit (1) is deployed using a spatial topology optimization algorithm to form a hemispherical monitoring network at the hydraulic station of the stamping machine, with equally spaced cross-intersection monitoring lines set at the welding station and an axial gradient distribution sensor installed on the main shaft of the cutting machine; The sensor group performs self-calibration during the first trouble-free operation phase of the equipment. By synchronously collecting raw data on dust concentration, vibration energy, and current fluctuation rate, and eliminating abnormal values ​​caused by the impact of equipment start-up and shutdown, the valid data is stored in the dust-vibration-current baseline comparison table, and a sliding window mechanism is used to calculate the dust concentration change rate to generate a dynamic dust accumulation gradient.

3. The intelligent predictive maintenance system for hardware tool production equipment according to claim 2, characterized in that: The trouble-free operation stage is divided into three stages: running-in period, stabilization period and aging verification period: During the run-in period, data on the correlation between dust concentration and vibration energy of the equipment under typical light-load conditions is collected to establish basic operating characteristics. During the stabilization period, data is extended to the normal production load range to obtain a dynamic mapping relationship between dust concentration and current fluctuation rate. During the aging verification period, sudden changes in dust concentration under extreme conditions are simulated to record the nonlinear response characteristics of vibration energy. After the three-stage data are jointly analyzed in the time domain and frequency domain, a three-dimensional baseline comparison table with adaptability to working conditions is generated. The table stores the correlation thresholds between dust concentration and vibration and current fluctuations under different operating conditions.

4. The intelligent predictive maintenance system for hardware tool production equipment according to claim 1, characterized in that: The extraction of the vibration energy peak adopts an improved time-frequency domain hybrid analysis method: The original vibration signal is decomposed by wavelet packets to separate the metal friction characteristic components in the high-frequency band. The root mean square value of the signal in this frequency band within the preset time window is calculated as the vibration energy benchmark. The instantaneous gradient is calculated based on the energy difference between adjacent windows. When the dust accumulation gradient exceeds the preset warning value, the high-frequency band energy tracking algorithm is activated to capture the bearing micro-vibration pulse waveform caused by dust intrusion.

5. The intelligent predictive maintenance system for hardware tool production equipment according to claim 4, characterized in that: The calculation of the current fluctuation rate includes a hardware compensation mechanism: When the dust sensor detects that the concentration exceeds the preset dust concentration threshold, the current ripple detection module starts the anti-interference mode to filter out the periodic harmonic components caused by fluctuations in the workshop power grid; The standard deviation of the effective current value in the recent preset time period is calculated using the moving average algorithm and compared with the historical data of the same dust concentration range in the baseline comparison table to generate a volatility deviation index in percentage form.

6. The intelligent predictive maintenance system for hardware tool production equipment according to claim 1, characterized in that: The three-level dust hazard classification introduces spatiotemporal correlation analysis technology: Low risk level determination: When the real-time dust concentration data is lower than the upper limit of the safety threshold marked in the dust-vibration-current baseline comparison table, and the change in the dust accumulation gradient over three consecutive monitoring cycles does not exceed the historical average fluctuation range recorded in the baseline comparison table, the current threshold remains unchanged; Medium-risk level determination: When the real-time dust concentration data exceeds the upper limit of the safety threshold but does not reach the emergency shutdown threshold in the baseline comparison table, or when the dust accumulation gradient shows a single-cycle mutation amplitude exceeding the historical average fluctuation range, the threshold tightening mechanism is activated to lower the determination thresholds of the vibration energy peak and current fluctuation rate; High-risk level determination: When the real-time dust concentration data exceeds the emergency shutdown threshold for a continuous preset period of time and the following conditions are met simultaneously, the dust high-risk state is triggered: (a) The current vibration energy peak exceeds the upper limit of the vibration threshold corresponding to the dust concentration range in the baseline comparison table; (b) the deviation percentage between the current fluctuation rate and the baseline value reaches or exceeds the preset deviation percentage of the recent maximum fluctuation value recorded by the dust feature library construction unit (2); (c) The dust accumulation gradient shows a continuous upward trend and does not trigger a response from the dust removal equipment.

7. The intelligent predictive maintenance system for hardware tool production equipment according to claim 1, characterized in that: The dynamic adjustment threshold adopts the device health feedback closed loop: The initial threshold is set based on the statistical average of the baseline comparison table. When the cumulative operating time of the equipment reaches the maintenance period, the threshold is positively compensated. When it is detected that the dust accumulation gradient enters a high-risk level, the threshold negative compensation algorithm is activated to adjust the threshold, while limiting the single adjustment amplitude according to the baseline value.

8. The intelligent predictive maintenance system for hardware tool production equipment according to claim 7, characterized in that: The calculation of the comprehensive dust impact index introduces the equipment type weight factor: The punching machine uses vibration energy deviation as the dominant factor, the welding equipment strengthens the contribution of current fluctuation, and the cutting machine introduces the spindle working condition attenuation coefficient; The calculation process performs an asymmetric fusion of dust real-time data and historical status based on a rolling time window, in which recent data has a dominant weight and historical data has an auxiliary weight, and finally outputs a standardized dust risk index.

9. The intelligent predictive maintenance system for hardware tool production equipment according to claim 8, characterized in that: The deviation comparison between the vibration energy peak and the baseline value is verified by multi-band coupling: The vibration signal is decomposed into the fundamental frequency band, the friction characteristic band and the dust impact band, and the deviation rate of the energy of each frequency band from the baseline value is calculated respectively; When the dust concentration reaches a high-risk level, the number of transient impacts in the dust impact section is monitored. If a pulse waveform with an amplitude exceeding the baseline value safety range appears more than n times within a preset time, it is directly determined to be a high-risk dust state.

10. The intelligent predictive maintenance system for hardware tool production equipment according to claim 9, characterized in that: The hierarchical control instructions include equipment process compensation parameters: Based on the hydraulic pressure attenuation coefficient output by the stamping machine hydraulic station, the main cylinder pressure is reduced according to the dust risk index, and a shielding gas flow compensation curve is generated for the welding equipment. Among them, the gas flow increase ratio is positively correlated with the current fluctuation rate. All instructions are virtually verified through the equipment digital twin model and then written into the PLC controller.

Images