Machine tool oil mist centralized management and control intelligent purification method and system

Abstract

The invention discloses a machine tool oil mist centralized management and control intelligent purification method and system, and relates to the technical field of oil mist management and control purification. State factors from different physical sources are subjected to normalization processing, the system breaks through a data dimension gap between sub-modules, and the management and control accuracy of the machine tool oil mist centralized management and control intelligent purification system is improved. The performance degradation behavior, the disturbance flow state behavior and the pollution residue behavior are uniformly incorporated into the same regulation and control intensity index system for the first time, and a uniform logic reference is provided for strategy evaluation. Different regulation and control intervals such as a normal area, a steady-state regulation and control area, a linkage control area and a warning response area correspond to different coping strategies respectively, abnormal data points caused by interference fluctuation, accidental jump and the like in oil mist parameter collection are effectively removed in a combined mode of a sliding window and median filtering, and the accuracy of oil mist parameter collection is improved. And it is ensured that data entering a subsequent calculation model has high stability and representativeness. The mechanism directly solves the common core problem of'high-noise data misleading judgment 'in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of oil mist control and purification, and specifically to a method and system for centralized control and intelligent purification of machine tool oil mist. Background Art

[0002] With the widespread use of CNC machine tools in mass precision manufacturing, processes such as high-speed cutting, lubrication spray, and coolant splash inevitably generate large amounts of oil-containing microparticles, or oil mist. This oil mist rapidly spreads within enclosed workshops, significantly impacting operator health, equipment safety, and the stability of the production environment. Especially in industrial scenarios where multiple machine tools are densely deployed and share a centralized purification system, dynamic control of residual oil mist and intelligent purification strategies have become core technical challenges, creating an urgent need for system-level solutions with multi-source sensing, closed-loop feedback loops, and intelligent regulation capabilities.

[0003] In most current factories, the oil mist purification systems deployed mostly use the traditional control method of fixed air volume + periodic maintenance + unified filter life estimation. Although this method has certain practicality in small-scale, single-station scenarios, it has serious shortcomings in scenarios with multi-station collaborative processing, uneven oil mist release, and frequent changes in process rhythm. First, the purification system cannot perceive the residual oil mist at the end of each station in real time, and therefore cannot adequately identify systemic failure signs such as filter load saturation, wind speed disturbance, and wind pressure backflow;

[0004] These technical deficiencies can lead to a series of adverse effects in actual operation. Because the amount of oil mist released during machine tool operation is highly correlated with the process steps, cutting speed, and coolant formulation, its residual concentration often exhibits "periodic spikes" and "uneven distribution across workstations." If the oil mist concentration in a particular workstation consistently exceeds the set threshold and the system fails to detect and respond promptly, oil mist can easily backflow into the air duct or adjacent workstations, increasing pollution levels throughout the entire workshop and making the source difficult to locate. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the present invention provides a method and system for centralized control and intelligent purification of machine tool oil mist to solve the problems mentioned in the background technology.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a machine tool oil mist centralized control intelligent purification system, including an oil mist residual parameter acquisition and processing module, a purification deviation identification and risk location module, an air duct abnormal coupling identification module, a filter load identification and attenuation assessment module, a state control intensity calculation and strategy generation module, and an instruction execution and closed-loop state memory module;

[0007] The oil mist residual parameter acquisition and processing module collects oil mist data through the sensor installed at the end of the air duct, fits it into the initial data group CW, and performs preprocessing to obtain the oil mist data set YW;

[0008] The purification deviation identification and risk location module extracts features from the oil mist dataset YW, extracts the current residual concentration, and compares it with the preset standard oil pollution threshold Tc to obtain the purification deviation coefficient Hpx and determine the purification status of the oil pollution collection point;

[0009] The duct abnormal coupling identification module analyzes the coordinated change trend of the duct static pressure Pa and the duct wind speed Sv according to the obtained purification status, obtains the non-steady-state coupling coefficient FUx, and identifies the local disturbance source and recirculation area;

[0010] The filter load identification and attenuation evaluation module analyzes the filter pressure difference Da and the oil mist volume Vu processed by the filter in the oil mist dataset YW to obtain the filter performance attenuation index SZw;

[0011] The state control intensity calculation and strategy generation module summarizes the obtained filter performance attenuation index SZw, non-steady-state coupling coefficient FUx and purification deviation coefficient Hpx, obtains the comprehensive control intensity index ZHs, and outputs the control strategy AC;

[0012] The instruction execution and closed-loop state memory module performs specific control actions according to the acquired control strategy AC, and dynamically adjusts the standard oil pollution threshold Tc according to the control results.

[0013] Preferably, the oil mist residual parameter acquisition and processing module comprises a multi-source oil mist parameter acquisition unit and an oil mist data pre-processing unit;

[0014] The multi-source oil mist parameter acquisition unit synchronously collects the oil mist data at the end of the air duct through sensors, including the mist residual concentration cRo, the static pressure Pa of the air duct, the wind speed Sv of the air duct, the filter pressure difference Da and the oil mist volume Vu processed by the filter, and fits them into the initial data set CW;

[0015] Among them, the residual concentration of fog cRo is collected by the laser scattering oil mist sensor installed at the air outlet at the end of the air duct;

[0016] The static pressure Pa of the air duct is collected and obtained by a micro differential pressure gas pressure sensor attached to the inner wall of the air duct;

[0017] The wind speed Sv in the air duct is acquired through a thermal wind speed sensor;

[0018] The filter pressure difference Da is collected and obtained by the pressure difference sensor installed in the cavity before and after the filter; specifically, it is obtained by the difference between the static pressure at the filter inlet and the static pressure at the filter outlet;

[0019] The oil mist volume Vu processed by the filter is collected and obtained by the cumulative flow sensor and the oil mist particle volume distribution estimation sensor; the oil mist volume Vu processed by the filter is obtained by the following formula:

[0020]

[0021] Where Vu(t) represents the volume of oil mist processed by the filter at time t, Sv(t) represents the wind speed in the duct at time t, Aduc represents the cross-sectional area of the duct, cRo(t) represents the residual concentration of mist at time t, and d represents the integral sign.

[0022] The oil mist data pre-processing unit performs abnormal elimination and standardization processing on the acquired initial data group CW to obtain a dimensionless oil mist data set YW for calculation;

[0023] Abnormal elimination is performed by using sliding window and median filtering methods to remove abnormal data in the initial data set CW;

[0024] Standardization processing is performed on the data in the initial data set CW by using the standard scaling method to obtain the oil mist data set YW;

[0025] The oil mist dataset YW is obtained by the following formula:

[0026]

[0027] Where YWp represents the p-th data item in the oil mist dataset YW, CWp represents the p-th data item in the initial data group CW, μCWp represents the mean of the p-th data item in the initial data group CW, and σCWp represents the standard deviation of the p-th data item in the initial data group CW.

[0028] Preferably, the purification deviation identification and risk positioning module includes a concentration feature extraction unit and a purification deviation coefficient construction unit;

[0029] The concentration feature extraction unit extracts the oil mist concentration features YcRo of all acquisition points at the current time point from the oil mist data set YW, and establishes the time evolution trend function YC;

[0030] The time evolution trend function YC is obtained by the following formula:

[0031] YC(t)={YcRo,1(t),YcRo,2(t),...,YcRo,n(t)};

[0032] Where YC(t) represents the time evolution trend function at time t, and n represents the total number of oil mist sampling points;

[0033] The concentration trend value LC is obtained by constructing a concentration smoothing sequence using an exponential sliding filter;

[0034] The concentration trend value LC is obtained by the following formula:

[0035] LCi(t)=α*YcRo,i(t)+(1-α)*LCi(t-1);

[0036] Where LCi(t) represents the concentration trend value of the i-th sampling point at time t, α represents the sliding filter coefficient, YcRo,i(t) represents the oil mist concentration characteristic of the i-th sampling point at time t, and LCi(t-1) represents the concentration trend value of the i-th sampling point at time t.

[0037] Preferably, the purification deviation coefficient construction unit compares the obtained concentration trend value LC with the standard oil pollution threshold value Tc to calculate and obtain the purification deviation coefficient Hpx;

[0038] The purification deviation coefficient Hpx is obtained by the following formula:

[0039]

[0040] Where, bc represents the risk magnification factor;

[0041] By obtaining the purification deviation coefficient Hpx, the purification failure risk index R is calculated and the purification status of the oil collection point is judged by the purification failure risk index R;

[0042] The purification failure risk index R is obtained by the following formula:

[0043]

[0044] In the formula, e represents a constant, and k represents the sensitivity factor that adjusts the steepness of the curve;

[0045] The purification status is obtained by matching:

[0046] When 0<purification failure risk index R<0.3, it indicates the first purification state and the purification is normal;

[0047] When 0.3≤purification failure risk index R<0.6, it indicates the second purification state, fluctuations occur, and observation is required;

[0048] When 0.6≤purification failure risk index R<1.0, it indicates the third purification state, purification is abnormal, and intervention is required.

[0049] Preferably, the air duct abnormal coupling identification module includes a local disturbance activation trigger unit and a non-steady-state coupling index calculation unit;

[0050] The local disturbance activation trigger unit analyzes the purification failure risk index R. When a purification anomaly occurs, the collection point where the purification anomaly occurs is screened and marked, and the duct number corresponding to the collection point is used as the core area for wind pressure and wind speed analysis. The duct static pressure Pa and duct wind speed Sv in the core area are synchronously paired.

[0051] The unsteady coupling index calculation unit differentiates and multiplies the paired duct static pressure Pa and duct wind speed Sv, and integrates them within a fixed period T to obtain the unsteady coupling coefficient FUx;

[0052] The unsteady coupling coefficient FUx is obtained by the following formula:

[0053]

[0054] Where Pa(t) represents the static pressure of the duct at time t, Sv(t) represents the wind speed of the duct at time t, and d represents the derivative symbol;

[0055] Compare the obtained non-steady-state coupling coefficient FUx with the preset abnormality identification threshold Tfu to determine the local disturbance state;

[0056] The local disturbance state is obtained by matching:

[0057] When the non-steady-state coupling coefficient FUx ≤ the abnormal identification threshold Tfu, it indicates that the local disturbance state is normal;

[0058] When the non-steady-state coupling coefficient FUx is greater than the abnormal identification threshold Tfu, it indicates that a local disturbance has occurred and a recirculation zone exists.

[0059] Preferably, the filter load identification and attenuation assessment module includes a filter load state feature extraction unit and a performance attenuation index construction unit;

[0060] The filter load state feature extraction unit takes the filter pressure difference Da and the oil mist volume Vu processed by the filter as the core input to construct the pressure difference growth factor Φ D and processing volume growth factor Φ V ;

[0061] Pressure difference growth factor Φ D Obtained by the ratio of the filter pressure difference Da(t) at time t to the initial design pressure difference Do of the filter;

[0062] Processing volume growth factor Φ V Obtained by the ratio of the oil mist volume Vu(t) processed by the filter at time t to the rated processing volume Vo of the filter;

[0063] The pressure difference growth factor Φ obtained by the performance degradation index construction unit D and processing volume growth factor ΦV Perform fusion and calculate the filter performance attenuation index SZw;

[0064] The filter performance degradation index SZw is obtained by the following formula:

[0065] SZw=log(Φ D *Φ V +ES);

[0066] Where, log represents the logarithmic function, ES represents a non-zero constant;

[0067] The filter performance attenuation index SZw is obtained to identify the filter status and determine whether the purification capacity has decreased;

[0068] The filter status is obtained by matching:

[0069] When 0<filter performance attenuation index SZw<0.3, it indicates that the first filter is in normal state; the filter working resistance is normal and has not reached the warning line, and there is still sufficient purification margin;

[0070] When 0.3≤filter performance attenuation index SZw<0.5, it indicates the second filter state, which is slightly attenuated; the filter is close to the performance limit, prompting planned replacement and suitable for scheduling;

[0071] When 0.5≤Filter Performance Degradation Index SZw<0.8, it indicates the third filter state, critical state; the filter performance has significantly decreased, it is recommended to stop part of the purification section and arrange for replacement as soon as possible;

[0072] When 0.8≤filter performance attenuation index SZw<1.0, it indicates the fourth filter state, overload state; the filter is seriously clogged, which may cause system pressure backflow and reflux pollution, and should be immediately decommissioned and replaced.

[0073] Preferably, the state control intensity calculation and strategy generation module includes a control intensity index calculation unit and a control strategy generation unit;

[0074] The control intensity index calculation unit normalizes the obtained filter performance attenuation index SZw, non-steady-state coupling coefficient FUx, and purification deviation coefficient Hpx, eliminates the dimensions of different parameters, unifies them into the same dimension, obtains the normalized filter performance attenuation index gSZw, the normalized non-steady-state coupling coefficient gFUx, and the normalized purification deviation coefficient gHpx, and fuses them to obtain the comprehensive control intensity index ZHs;

[0075] The comprehensive regulation intensity index ZHs is obtained by the following formula:

[0076] ZHs=β1*gSZw+β2*gFUx+β3*gHpx;

[0077] Where β1, β2, and β3 represent the preset weight values of the normalized filter performance attenuation index gSZw, the normalized non-steady-state coupling coefficient gFUx, and the normalized purification deviation coefficient gHpx, respectively, and β1+β2+β3≤1.

[0078] Preferably, the control strategy generation unit compares the obtained comprehensive control intensity index ZHs with a preset control threshold Tzh to obtain a control strategy AC;

[0079] Control strategy AC is obtained by matching in the following ways:

[0080] When the control intensity index ZHs is less than the control threshold Tzh*0.5, it indicates the normal zone, and the control strategy AC is: no action and small ventilation adjustment;

[0081] When the control threshold Tzh*0.5≤control intensity index ZHs<control threshold Tzh, it indicates the steady-state control zone, and the control strategy AC: air volume fine-tuning and wind direction bias strategy;

[0082] When the control threshold Tzh ≤ the control intensity index ZHs < the control threshold Tzh*1.5, it indicates the linkage control area, and the control strategy AC: filter replacement warning and path switching;

[0083] When the control threshold Tzh*1.5≤control intensity index ZHs<control threshold Tzh*2.0, it indicates the warning response area, and the control strategy AC: deactivate some air sections, force filter replacement prompts and isolate abnormal points.

[0084] Preferably, the instruction execution and closed-loop state memory module includes a control instruction execution unit and an adaptive adjustment unit;

[0085] The control instruction execution unit interprets the acquired control strategy AC as a device-level control instruction, driving the physical device to operate, including adjusting the air valve angle, switching the purification path, and lighting the filter indicator light. It also synchronously collects the effect information after the control execution through the feedback loop.

[0086] After executing the control strategy AC, the adaptive adjustment unit conducts a trend assessment on the control execution results. When the purification deviation coefficient Hpx continues to decrease and the control intensity index ZHs is less than the control threshold Tzh, the dynamic downward adjustment mechanism of the standard oil pollution threshold Tc is triggered to obtain a new oil pollution threshold Tc.

[0087] According to the purification deviation coefficient Hpx, the purification deviation change rate ΔHpx of N consecutive cycles is calculated;

[0088] The purification deviation change rate ΔHpx is obtained by the following formula:

[0089]

[0090] Where N is the number of time window periods, Q is the sum index variable, which accumulates from 1 to N and is used to retrieve the historical data series. Hpx(tQ) is the purification deviation coefficient at time tQ, and Hpx(t-Q+1) is the purification deviation coefficient at time t-Q+1. It is used to calculate the deviation change for each time period by subtracting it from the previous period.

[0091] The new oil pollution threshold Tc is obtained by the following formula:

[0092] Tc(t+1)=Tc(t)*(1-λ*ΔHpx);

[0093] Where Tc(t+1) represents the standard oil contamination threshold at time t+1, Tc(t) represents the standard oil contamination threshold at time t, and λ represents the adjustment sensitivity coefficient.

[0094] The present invention also provides a method for centralized control and intelligent purification of machine tool oil mist, comprising the following steps:

[0095] Step 1: The oil mist residual parameter acquisition and processing module collects oil mist data through the sensor installed at the end of the air duct, fits it into the initial data set CW, and performs preprocessing to obtain the oil mist data set YW;

[0096] Step 2: The purification deviation identification and risk location module performs feature extraction on the oil mist dataset YW, extracts the current residual concentration, and compares it with the preset standard oil pollution threshold Tc to obtain the purification deviation coefficient Hpx and determine the purification status of the oil pollution collection point;

[0097] Step 3: The duct abnormal coupling identification module analyzes the coordinated change trend of the duct static pressure Pa and the duct wind speed Sv based on the obtained purification status, obtains the non-steady-state coupling coefficient FUx, and identifies the local disturbance source and recirculation area;

[0098] Step 4: The filter load identification and attenuation evaluation module analyzes the filter pressure difference Da and the oil mist volume Vu processed by the filter in the oil mist dataset YW to obtain the filter performance attenuation index SZw;

[0099] Step 5: The state control intensity calculation and strategy generation module summarizes the obtained filter performance attenuation index SZw, non-steady-state coupling coefficient FUx and purification deviation coefficient Hpx to obtain the comprehensive control intensity index ZHs and output the control strategy AC;

[0100] Step 6: The instruction execution and closed-loop state memory module performs specific control actions according to the acquired control strategy AC, and dynamically adjusts the standard oil pollution threshold Tc according to the control results.

[0101] The present invention provides a method and system for centralized control and intelligent purification of machine tool oil mist, which has the following beneficial effects:

[0102] (1) When the system is running, a control intensity index is constructed and a response strategy set is generated. It can perform precise intervention according to the current system operation status, such as automatically adjusting the air volume, switching the purification path, and triggering early warning prompts, thus realizing the leap from "manual experience-based" to "data-driven" control. Based on the changing trend of continuous purification deviation, the system can automatically and appropriately lower the threshold Tc under the conditions of stable system operation and continuous deviation reduction, thereby enhancing the accuracy of the purification standard and the system sensitivity, and realizing the adaptive optimization evolution of the control standard.

[0103] (2) By combining the sliding window with the median filter, abnormal data points caused by interference fluctuations, occasional jumps, etc. in the oil mist parameter collection are effectively removed, ensuring that the data entering the subsequent calculation model is highly stable and representative. This mechanism directly solves the core problem of "high-noise data misleading judgment" that is prevalent in existing technologies. Through the preliminary preprocessing process, the system can directly transmit reliable and uniformly structured data sets to downstream purification deviation identification, air duct coupling judgment, filter attenuation analysis and other modules, forming a system-level upstream and downstream decoupling, and improving the overall response speed and robustness of the system.

[0104] (3) By normalizing the state factors from different physical sources, the system breaks down the data dimension barriers between the submodules and for the first time integrates performance degradation behavior, disturbed flow behavior, and pollution residual behavior into the same control intensity index system, providing a unified logical benchmark for strategy evaluation. Different control intervals, such as normal zone, steady-state control zone, linkage control zone, and warning response zone, correspond to different response strategies, such as fine-tuning, wind direction bias, filter warning, and equipment isolation. Compared with the traditional "one threshold + single response" rigid strategy, this mechanism can provide a progressive, multi-level control response according to the system operation status, which not only ensures stability, but also reduces resource waste and improves the overall operation economy and safety.

[0105] (4) By dividing the filter performance attenuation index SZw into four stages, the system can present the filter from normal, mild, critical to overload, providing maintenance personnel with concrete and graded treatment suggestions, thereby improving the scientific nature and planning of maintenance management. By uniformly calculating the comprehensive control intensity index ZHs based on the filter status, airflow coupling anomaly, and purification deviation, the system achieves cross-modal fusion and intelligent trade-off judgment, enabling the control strategy to match the actual degree of anomaly, evolving from a "single-point response" to a system-linked intervention. BRIEF DESCRIPTION OF THE DRAWINGS

[0106] Figure 1 This is a block diagram and flow chart of the intelligent purification system for centralized control of oil mist in machine tools according to the present invention;

[0107] Figure 2 Schematic diagram of the steps of the intelligent purification method for centralized control of machine tool oil mist according to the present invention;

[0108] Figure 3 This is a schematic diagram of the control strategy framework flow of the present invention;

[0109] Figure 4 It is a line graph of the comprehensive regulation intensity index of the present invention. DETAILED DESCRIPTION

[0110] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0111] Example 1

[0112] The present invention provides a centralized control and intelligent purification system for machine tool oil mist, please refer to Figures 1 to 4 The system includes an oil mist residual parameter acquisition and processing module, a purification deviation identification and risk positioning module, an air duct abnormal coupling identification module, a filter load identification and attenuation evaluation module, a state control intensity calculation and strategy generation module, and an instruction execution and closed-loop state memory module.

[0113] The oil mist residual parameter acquisition and processing module collects the oil mist data through the sensor installed at the end of the air duct, fits it into the initial data group CW, and performs preprocessing to obtain the oil mist data set YW.

[0114] The purification deviation identification and risk positioning module extracts features from the oil mist dataset YW, extracts the current residual concentration, and compares it with the preset standard oil pollution threshold Tc to obtain the purification deviation coefficient Hpx and judge the purification status of the oil pollution collection point.

[0115] The duct abnormal coupling identification module analyzes the coordinated change trend of the duct static pressure Pa and the duct wind speed Sv according to the obtained purification status, obtains the non-steady-state coupling coefficient FUx, and identifies the local disturbance source and recirculation area.

[0116] The filter load identification and attenuation evaluation module analyzes the filter pressure difference Da and the oil mist volume Vu processed by the filter in the oil mist dataset YW to obtain the filter performance attenuation index SZw.

[0117] The state control intensity calculation and strategy generation module summarizes the obtained filter performance attenuation index SZw, non-steady-state coupling coefficient FUx and purification deviation coefficient Hpx, obtains the comprehensive control intensity index ZHs, and outputs the control strategy AC.

[0118] The instruction execution and closed-loop state memory module performs specific control actions according to the acquired control strategy AC, and dynamically adjusts the standard oil pollution threshold Tc according to the control results.

[0119] In this embodiment, the system, for the first time, integrates the parallel collection of multiple physical parameters, including oil mist residue, static pressure, wind speed, and filter load, at the end of the air duct. These parameters are then converted into a standardized dataset, eliminating the reliance on single concentration data for purification effect analysis. This provides enhanced data integrity and adaptability to operating conditions. By comparing residual oil mist concentration with a preset threshold in real time and constructing a coefficient of deviation determination model, the system dynamically identifies the effectiveness of purification, shifting from a crude determination of "whether the standard is exceeded" to a more refined identification of "changes in purification trends."

[0120] When the purification status is abnormal, the system can further retrospectively analyze the coordinated evolution of static pressure and wind speed in the air duct, identifying local disturbance sources or areas of airflow recirculation. This solves the problem of existing systems being unable to visualize and control internal duct disturbances. The system combines filter pressure differential with processing volume to construct a performance degradation index, supporting dynamic assessment of filter load status. This shifts filter maintenance from scheduled to on-demand replacement, reducing consumables waste and improving system cost-effectiveness and reliability.

[0121] The system integrates three core indicators to construct a control intensity index and generate a set of response strategies. This allows for precise intervention based on the current system operating status, such as automatically adjusting air volume, switching purification paths, and triggering early warning prompts. This represents a leap from manual, empirical control to data-driven control. Based on the changing trend of continuous purification deviations, the system automatically and appropriately lowers the threshold Tc when system operation is stable and deviations continue to decrease, thereby enhancing the accuracy of the purification standard and system sensitivity, achieving adaptive optimization and evolution of the control standard.

[0122] Example 2

[0123] This embodiment is explained in Example 1, please refer to Figure 1 and Figure 3 Specifically, the oil mist residual parameter acquisition and processing module includes a multi-source oil mist parameter acquisition unit and an oil mist data preprocessing unit.

[0124] The multi-source oil mist parameter acquisition unit synchronously collects the oil mist data at the end of the air duct through sensors, including the mist residual concentration cRo, the static pressure Pa of the air duct, the wind speed Sv of the air duct, the filter pressure difference Da and the oil mist volume Vu processed by the filter, and fits them into the initial data set CW.

[0125] The residual mist concentration cRo is collected by a laser scattering oil mist sensor installed at the air outlet at the end of the duct. The static pressure Pa of the duct is collected by a micro-differential pressure gas pressure sensor attached to the inner wall of the duct. The duct wind speed Sv is collected by a thermal wind speed sensor. The filter pressure difference Da is collected by pressure difference sensors installed in the cavities before and after the filter. The oil mist volume Vu processed by the filter is collected by a cumulative flow sensor and an oil mist particle volume distribution estimation sensor. The oil mist data preprocessing unit removes anomalies and normalizes the initial data set CW to obtain a dimensionless oil mist dataset YW for calculation. Abnormal data in the initial data set CW is removed by using a sliding window and median filtering method. Normalization is performed by processing the data in the initial data set CW using a standard scaling method to obtain the oil mist dataset YW.

[0126] The oil mist dataset YW is obtained by the following formula:

[0127]

[0128] Where YWp represents the p-th data item in the oil mist dataset YW, CWp represents the p-th data item in the initial data group CW, μCWp represents the mean of the p-th data item in the initial data group CW, and σCWp represents the standard deviation of the p-th data item in the initial data group CW.

[0129] Through systematic deployment at key points such as the end of the air duct and before and after the filter, this system achieves the first integrated and synchronized collection of multiple key indicators, including residual fog concentration cRo, duct static pressure Pa, duct velocity Sv, filter pressure difference Da, and oil mist volume Vu processed by the filter. Compared to traditional single-point detection methods that only monitor concentration, this system significantly enhances its comprehensive perception of the operating environment.

[0130] By uniformly fitting the initial data set CW, the heterogeneous data sources are uniformly encoded and encapsulated, ensuring the logical consistency and simplicity of processing in subsequent feature extraction and indicator construction, overcoming the limitations of traditional systems where different parameters are collected independently and data integration is difficult. By combining sliding window and median filtering, abnormal data points caused by interference fluctuations, occasional jumps, etc. in oil mist parameter collection are effectively removed, ensuring that the data entering the subsequent calculation model is highly stable and representative. This mechanism directly solves the core problem of "high-noise data misleading judgment" that is prevalent in existing technologies. Through the preliminary preprocessing process, the system can directly transmit reliable and uniformly structured data sets to downstream purification deviation identification, air duct coupling judgment, filter attenuation analysis and other modules, forming system-level upstream and downstream decoupling, and improving the overall response speed and robustness of the system.

[0131] The YW dataset constructed in this example already possesses a standardized data structure suitable for subsequent dynamic assessment, risk modeling, and feedback control. This provides the starting conditions for information integrity and algorithm adaptability in the closed-loop purification-monitoring-correction mechanism, which is unattainable using traditional data collection logic. Data standardization lays the foundation for future expansion of the system's machine learning models and the construction of historical trend recognition mechanisms. This enables the system to not only "judge in the present" but also "remember the past" and "predict trends," significantly enhancing its potential for intelligent evolution.

[0132] Example 3

[0133] This embodiment is explained in Example 2, please refer to Figure 1 ,Specifically, the purification deviation identification and risk ,location module includes a concentration feature extraction unit and a purification ,deviation coefficient construction unit.

[0134] The concentration feature extraction unit extracts the oil mist concentration features YcRo of all acquisition points at the current time point from the oil mist dataset YW, and establishes a time evolution trend function YC.

[0135] The time evolution trend function YC is obtained by the following formula:

[0136] YC(t)={YcRo,1(t),YcRo,2(t),...,YcRo,n(t)};

[0137] Where YC(t) represents the time evolution trend function at time t, and n represents the total number of oil mist sampling points.

[0138] The concentration trend value LC is obtained by constructing a concentration smoothing sequence using an exponential sliding filter.

[0139] The concentration trend value LC is obtained by the following formula:

[0140] LCi(t)=α*YcRo,i(t)+(1-α)*LCi(t-1);

[0141] Where LCi(t) represents the concentration trend value of the i-th sampling point at time t, α represents the sliding filter coefficient, YcRo,i(t) represents the oil mist concentration characteristic of the i-th sampling point at time t, and LCi(t-1) represents the concentration trend value of the i-th sampling point at time t.

[0142] The purification deviation coefficient construction unit compares the obtained concentration trend value LC with the standard oil pollution threshold value Tc and calculates the purification deviation coefficient Hpx.

[0143] The purification deviation coefficient Hpx is obtained by the following formula:

[0144]

[0145] Where bc represents the risk magnification factor.

[0146] By obtaining the purification deviation coefficient Hpx, the purification failure risk index R is calculated, and the purification status of the oil collection point is judged by the purification failure risk index R.

[0147] The purification failure risk index R is obtained by the following formula:

[0148]

[0149] Where e is a constant and k is a sensitivity factor that adjusts the steepness of the curve.

[0150] The purification status is obtained by matching:

[0151] When 0<purification failure risk index R<0.3, it indicates the first purification state and the purification is normal;

[0152] When 0.3≤purification failure risk index R<0.6, it indicates the second purification state, fluctuations occur, and observation is required;

[0153] When 0.6≤purification failure risk index R<1.0, it indicates the third purification state, purification is abnormal, and intervention is required.

[0154] In this embodiment, an exponential sliding filtering mechanism is introduced through the concentration feature extraction unit to construct a continuous concentration trend function, so that the system can identify "whether the purification quality is improving, stabilizing or deteriorating", realizing continuous perception of dynamic trends and significantly improving the system's predictive and stability recognition capabilities.

[0155] This embodiment no longer relies solely on static judgments of "exceeding / not exceeding" based on whether the concentration is above a threshold. Instead, it constructs a purification deviation coefficient, Hpx, to compare the current concentration trend with a set standard oil pollution threshold. It also introduces a risk amplification factor to address the importance of sensitive areas or key sampling points, enabling the system to identify regional differences and significantly improving the accuracy and specificity of judgments. Because this module outputs structured data with levels, trends, and risk factors, rather than a single threshold judgment, it facilitates refined control and decentralized regulation in downstream modules, making the overall system's response logic more intelligent and collaborative.

[0156] Example 4

[0157] This embodiment is explained in Example 3, please refer to Figure 3 ,Specifically, the air duct abnormal coupling identification module includes a ,local disturbance activation trigger unit and a non-steady-state coupling index calculation unit.

[0158] The local disturbance activation trigger unit analyzes the purification failure risk index R. When a purification anomaly occurs, the collection point where the purification anomaly occurs is screened and marked, and the duct number corresponding to the collection point is used as the core area for wind pressure and wind speed analysis. The duct static pressure Pa and duct wind speed Sv in the core area are synchronously paired.

[0159] The unsteady coupling index calculation unit differentiates and multiplies the paired duct static pressure Pa and duct wind speed Sv, and integrates them within a fixed period T to obtain the unsteady coupling coefficient Fux.

[0160] The unsteady coupling coefficient FUx is obtained by the following formula:

[0161]

[0162] Where Pa(t) represents the static pressure of the duct at time t, Sv(t) represents the wind speed of the duct at time t, and d represents the derivative symbol.

[0163] The obtained non-steady-state coupling coefficient FUx is compared with the preset abnormality identification threshold Tfu to determine the local disturbance state.

[0164] The local disturbance state is obtained by matching:

[0165] When the non-steady-state coupling coefficient FUx ≤ the abnormal identification threshold Tfu, it indicates that the local disturbance state is normal;

[0166] When the non-steady-state coupling coefficient FUx is greater than the abnormal identification threshold Tfu, it indicates that a local disturbance has occurred and a recirculation zone exists.

[0167] The filter load identification and attenuation evaluation module includes a filter load state feature extraction unit and a performance attenuation index construction unit. The filter load state feature extraction unit takes the filter pressure difference Da and the oil mist volume Vu processed by the filter as the core inputs and constructs the pressure difference growth factor Φ D and processing volume growth factor Φ V . Pressure difference growth factor Φ D It is obtained by the ratio of the filter pressure difference Da(t) at time t to the initial design filter pressure difference Do.

[0168] Processing volume growth factor Φ V The ratio of the oil mist volume Vu(t) processed by the filter at time t to the rated processing volume Vo of the filter is obtained. The performance degradation index construction unit is used to obtain the pressure difference growth factor Φ D and processing volume growth factor Φ V The fusion is performed to calculate the filter performance attenuation index SZw.

[0169] The filter performance degradation index SZw is obtained by the following formula:

[0170] SZw=log(Φ D *Φ V +ES);

[0171] Where log represents the logarithmic function, and ES represents a non-zero constant.

[0172] The filter status is identified by the filter performance degradation index SZw, and whether the purification capacity has decreased is determined. The filter status is obtained by matching in the following ways:

[0173] When 0<filter performance attenuation index SZw<0.3, it indicates the first filter state, normal state;

[0174] When 0.3≤filter performance attenuation index SZw<0.5, it indicates the second filter state, slight attenuation;

[0175] When 0.5≤filter performance attenuation index SZw<0.8, it indicates the third filter state, critical state;

[0176] When 0.8≤filter performance attenuation index SZw<1.0, it indicates the fourth filter state, overload state.

[0177] This embodiment introduces a local disturbance activation trigger unit, which uses the purification failure risk index as a prerequisite criterion and limits the identification range to the air duct area corresponding to the abnormal sampling point, avoiding invalid full-area scanning, significantly improving the system's response focusing capability and energy efficiency, and also solving the problem of the inability to locate air duct anomalies and the large waste of control resources in traditional systems. Through the differential coupling calculation of the air duct static pressure and wind speed, the system can identify the airflow disturbance intensity and change trend. This co-evolution relationship is a hidden failure factor that traditional purification systems cannot perceive. This improved mechanism enables the system to have the ability to detect micro-flow field phenomena such as backflow, turbulence, and compression disturbances, significantly improving the early warning capability of abnormal air duct behavior.

[0178] The system no longer relies on simple air volume mutation values or single-variable jump thresholds. Instead, it integrates and analyzes the level of "disturbance accumulation energy" to more comprehensively determine whether there are backflow areas or disturbance sources. This quantifies and integrates disturbance behavior, logically analyzing the judgment, and improves the scientific and robust nature of system diagnosis. By integrating the pressure difference factor with the volume factor and incorporating logarithmic processing, the system can achieve a significant exponential amplification response before the filter attenuation approaches its limit, thereby issuing a critical state warning before the filter is completely blocked, effectively solving the problems of delayed replacement and contamination backflow.

[0179] The two major functional modules of this embodiment respectively perform quantitative analysis on the stability of the airflow structure and the degradation of the filter purification performance, and through a unified threshold judgment mechanism, form a closed-loop control pre-input, providing the state control module with an input basis with a stronger "background state understanding capability", so that the entire system has the ability to judge the cause-effect chain and proactively defend against risks. Whether it is airflow disturbance caused by rapid changes in wind speed or premature attenuation of the filter due to increased concentration load, this embodiment provides an indicator mechanism that can be dynamically perceived and accurately identified, so that the system has higher operational stability and self-regulation capabilities when facing the inevitable uneven load and unstable rhythm in the actual operation of the workshop.

[0180] Example 5

[0181] This embodiment is explained in Example 4. Please refer to Figure 1 and Figure 4 Specifically, the state control strength calculation and strategy generation module includes a control strength index calculation unit and a control strategy generation unit.

[0182] The control intensity index calculation unit normalizes the obtained filter performance attenuation index SZw, non-steady-state coupling coefficient FUx, and purification deviation coefficient Hpx, eliminating the dimensions of different parameters and unifying them into the same dimension to obtain the normalized filter performance attenuation index gSZw, the normalized non-steady-state coupling coefficient gFUx, and the normalized purification deviation coefficient gHpx, and fuses them to obtain the comprehensive control intensity index ZHs. The comprehensive control intensity index ZHs is obtained by the following formula:

[0183] ZHs=β1*gSZw+β2*gFUx+β3*gHpx;

[0184] Where β1, β2, and β3 represent the preset weight values of the normalized filter performance attenuation index gSZw, the normalized non-steady-state coupling coefficient gFUx, and the normalized purification deviation coefficient gHpx, respectively, and β1+β2+β3≤1.

[0185] Specific examples:

[0186] Table 1: Calculation table of comprehensive regulation intensity index;

[0187]

[0188]

[0189] The control strategy generation unit compares the obtained comprehensive control intensity index ZHs with the preset control threshold Tzh to obtain the control strategy AC. The control strategy AC is obtained by matching:

[0190] When the control intensity index ZHs is less than the control threshold Tzh*0.5, it indicates the normal zone, and the control strategy AC is: no action and small ventilation adjustment;

[0191] When the control threshold Tzh*0.5≤control intensity index ZHs<control threshold Tzh, it indicates the steady-state control zone, and the control strategy AC: air volume fine-tuning and wind direction bias strategy;

[0192] When the control threshold Tzh ≤ the control intensity index ZHs < the control threshold Tzh*1.5, it indicates the linkage control area, and the control strategy AC: filter replacement warning and path switching;

[0193] When the control threshold Tzh*1.5≤control intensity index ZHs<control threshold Tzh*2.0, it indicates the warning response area, and the control strategy AC: deactivate some air sections, force filter replacement prompts and isolate abnormal points.

[0194] The command execution and closed-loop state memory module includes a control command execution unit and an adaptive adjustment unit. The control command execution unit interprets the acquired control strategy AC as device-level control commands, driving physical device operations including damper angle adjustment, purification path switching, and filter indicator lighting. It also synchronously collects information about the effects of control execution through a feedback loop.

[0195] After executing the control strategy AC, the adaptive adjustment unit conducts a trend evaluation on the control execution results. When the purification deviation coefficient Hpx continues to decrease and the control intensity index ZHs is less than the control threshold Tzh, the dynamic downward adjustment mechanism of the standard oil pollution threshold Tc is triggered to obtain a new oil pollution threshold Tc.

[0196] Based on the purification deviation coefficient Hpx, calculate the purification deviation change rate ΔHpx for N consecutive cycles. The purification deviation change rate ΔHpx is obtained by the following formula:

[0197]

[0198] Where N represents the number of time window periods, Q represents the summation index variable, Hpx(tQ) represents the purification deviation coefficient at time tQ, and Hpx(t-Q+1) represents the purification deviation coefficient at time t-Q+1.

[0199] The new oil pollution threshold Tc is obtained by the following formula:

[0200] Tc(t+1)=Tc(t)*(1-λ*ΔHpx);

[0201] Where Tc(t+1) represents the standard oil contamination threshold at time t+1, Tc(t) represents the standard oil contamination threshold at time t, and λ represents the adjustment sensitivity coefficient.

[0202] In this embodiment, by normalizing state factors from different physical sources, the system breaks down the data dimension barriers between submodules and, for the first time, integrates performance degradation behavior, disturbed flow behavior, and residual pollution behavior into the same control intensity indicator system, providing a unified logical benchmark for strategy evaluation. Different control intervals, such as the normal zone, steady-state control zone, linkage control zone, and alert response zone, correspond to different response strategies, such as fine-tuning, wind direction bias, filter warning, and equipment isolation. Compared to the traditional "one threshold + single response" rigid strategy approach, this mechanism can provide a progressive, multi-level control response based on the system's operating status, ensuring stability while reducing resource waste and improving overall operational economy and safety.

[0203] The strategy generation unit is not limited to air volume control, but also includes various physical actions such as wind direction adjustment, path switching and filter replacement prompts, so that the system has the combined response capability to complex multi-source pollution problems, and can achieve precise response matching based on the changing trend of the control intensity index.

[0204] In traditional systems, the standard oil pollution threshold Tc is a fixed parameter and cannot be automatically adjusted according to environmental changes. It is very easy to have problems such as "the threshold is too high, resulting in purification failure" or "the threshold is too low, resulting in false alarms." This embodiment introduces a dynamic adjustment mechanism by evaluating the change rate of continuous purification deviations, so that the system can adaptively tighten standards and improve purification requirements when the deviation continues to improve, significantly improving subsequent purification accuracy and response sensitivity. Once the control strategy is executed, it is no longer "execution and completion", but is driven by feedback information to further judge the execution effect and affect the standard update. This mechanism changes the drawbacks of the traditional "control-result separation" and realizes a truly self-repairing, self-driving, and self-evolving purification control path.

[0205] The generation of the comprehensive control intensity index (ZHs) is based on the aggregation of multiple core upstream indicators and provides a driving force for the generation of action commands for downstream modules. This forms a complete system control closed loop: multi-factor input - centralized judgment - precise response - real-time feedback - parameter update, driving the system's transition from "passive monitoring" to "active evolution." The generated control strategy (AC) is not only responsive but also quantifiable, explainable, and predictable, facilitating problem tracing, event analysis, and decision-making intervention for technical personnel. This is a key step in building industrial-grade intelligent control systems.

[0206] Example 6

[0207] The present invention also provides a method for centralized control and intelligent purification of machine tool oil mist, please refer to Figure 2 Specifically, the method includes the following steps:

[0208] Step 1: The oil mist residual parameter acquisition and processing module collects oil mist data through the sensor installed at the end of the air duct, fits it into the initial data set CW, and performs preprocessing to obtain the oil mist data set YW;

[0209] Step 2: The purification deviation identification and risk location module performs feature extraction on the oil mist dataset YW, extracts the current residual concentration, and compares it with the preset standard oil pollution threshold Tc to obtain the purification deviation coefficient Hpx and determine the purification status of the oil pollution collection point;

[0210] Step 3: The duct abnormal coupling identification module analyzes the coordinated change trend of the duct static pressure Pa and the duct wind speed Sv based on the obtained purification status, obtains the non-steady-state coupling coefficient FUx, and identifies the local disturbance source and recirculation area;

[0211] Step 4: The filter load identification and attenuation evaluation module analyzes the filter pressure difference Da and the oil mist volume Vu processed by the filter in the oil mist dataset YW to obtain the filter performance attenuation index SZw;

[0212] Step 5: The state control intensity calculation and strategy generation module summarizes the obtained filter performance attenuation index SZw, non-steady-state coupling coefficient FUx and purification deviation coefficient Hpx to obtain the comprehensive control intensity index ZHs and output the control strategy AC;

[0213] Step 6: The instruction execution and closed-loop state memory module performs specific control actions according to the acquired control strategy AC, and dynamically adjusts the standard oil pollution threshold Tc according to the control results.

[0214] The first step of this approach is to simultaneously collect key parameters such as residual fog concentration, duct static pressure, wind speed, filter pressure difference, and oil mist volume through a combination of multiple sensors. This establishes a systematic and realistic initial dataset reflecting the on-site status, providing a solid data foundation for subsequent intelligent identification and strategy formulation. This addresses the challenges of traditional systems, which suffer from a single data source and ambiguous state identification. The introduction of an abnormal data elimination and standardization mechanism enables the system to extract core trend features even in high-noise conditions, ensuring the stability and representativeness of the data relied upon by subsequent modules and improving system robustness.

[0215] The purification deviation identification and risk location module no longer relies on fixed thresholds, but instead introduces a trend-based judgment mechanism. This allows for early identification of potential purification failures, resolving the drawback of relying solely on delayed alarm responses and enhancing the system's predictive control capabilities. When a purification anomaly is detected, the system automatically identifies the abnormal sampling point area and further analyzes the coordinated behavior of wind speed and static pressure in the corresponding air duct. This enables the first automatic identification of duct recirculation areas and disturbance sources, providing "causal traceability" for precise control.

[0216] This method no longer simply judges the operating status of the filter based on pressure difference or time, but instead integrates the actual filtration volume and resistance growth trend to construct an exponentially responding performance decay index, which improves the judgment sensitivity of the critical failure stage and helps to issue an effective warning before the filter enters the overload state.

[0217] By categorizing the filter performance degradation index (SZw) into four status intervals, the system presents the filter in four stages: normal, mild, critical, and overloaded. This provides maintenance personnel with concrete, graded recommendations for handling the issue, enhancing the scientific and planned nature of maintenance management. By integrating filter status, airflow coupling anomalies, and purification deviations into a comprehensive control intensity index (ZHs), the system achieves cross-modal fusion and intelligent trade-offs, enabling control strategies to match the actual degree of anomaly, evolving from a "single-point response" approach to a system-wide, coordinated intervention approach.

[0218] Each control instruction corresponds to the feedback result analysis. When continuous purification improvement is found, the system automatically adjusts the judgment criteria, truly establishing a complete closed-loop process of "judgment-response-feedback-update". It has the characteristics of self-evolution and self-optimization, which is incomparable to traditional control logic.

[0219] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A machine tool oil mist centralized control intelligent purification system, characterized in that: It includes oil mist residual parameter collection and processing module, purification deviation identification and risk positioning module, air duct abnormal coupling identification module, filter load identification and attenuation assessment module, state control intensity calculation and strategy generation module, and instruction execution and closed-loop state memory module; The oil mist residual parameter acquisition and processing module collects oil mist data through the sensor installed at the end of the air duct, fits it into the initial data group CW, and performs preprocessing to obtain the oil mist data set YW; The purification deviation identification and risk location module extracts features from the oil mist dataset YW, extracts the current residual concentration, and compares it with the preset standard oil pollution threshold Tc to obtain the purification deviation coefficient Hpx and determine the purification status of the oil pollution collection point; The duct abnormal coupling identification module analyzes the coordinated change trend of the duct static pressure Pa and the duct wind speed Sv according to the obtained purification status, obtains the non-steady-state coupling coefficient FUx, and identifies the local disturbance source and recirculation area; The filter load identification and attenuation evaluation module analyzes the filter pressure difference Da and the oil mist volume Vu processed by the filter in the oil mist dataset YW to obtain the filter performance attenuation index SZw; The state control intensity calculation and strategy generation module summarizes the obtained filter performance attenuation index SZw, non-steady-state coupling coefficient FUx and purification deviation coefficient Hpx, obtains the comprehensive control intensity index ZHs, and outputs the control strategy AC; The instruction execution and closed-loop state memory module performs specific control actions according to the acquired control strategy AC, and dynamically adjusts the standard oil pollution threshold Tc according to the control results.

2. The machine tool oil mist centralized control intelligent purification system according to claim 1 is characterized in that: Oil mist residual parameter acquisition and processing module Multi-source oil mist parameter acquisition unit and oil mist data pre-processing unit; The multi-source oil mist parameter acquisition unit synchronously collects the oil mist data at the end of the air duct through sensors, including the mist residual concentration cRo, the static pressure Pa of the air duct, the wind speed Sv of the air duct, the filter pressure difference Da and the oil mist volume Vu processed by the filter, and fits them into the initial data set CW; Among them, the residual concentration of fog cRo is collected by the laser scattering oil mist sensor installed at the air outlet at the end of the air duct; The static pressure Pa of the air duct is collected and obtained by a micro differential pressure gas pressure sensor attached to the inner wall of the air duct; The wind speed Sv in the air duct is acquired through a thermal wind speed sensor; The filter pressure difference Da is collected and obtained through the pressure difference sensors installed in the cavity before and after the filter; The oil mist volume Vu processed by the filter is collected and obtained through the cumulative flow sensor and the oil mist particle volume distribution estimation sensor; The oil mist data pre-processing unit performs abnormal elimination and standardization processing on the acquired initial data group CW to obtain a dimensionless oil mist data set YW for calculation; Abnormal elimination is performed by using sliding window and median filtering methods to remove abnormal data in the initial data set CW; Standardization processing is performed on the data in the initial data set CW by using the standard scaling method to obtain the oil mist data set YW; The oil mist dataset YW is obtained by the following formula: Where YWp represents the p-th data item in the oil mist dataset YW, CWp represents the p-th data item in the initial data group CW, μCWp represents the mean of the p-th data item in the initial data group CW, and σCWp represents the standard deviation of the p-th data item in the initial data group CW.

3. The intelligent purification system for centralized control of machine tool oil mist according to claim 2 is characterized in that: The purification deviation identification and risk positioning module includes a concentration feature extraction unit and a purification deviation coefficient construction unit; The concentration feature extraction unit extracts the oil mist concentration features YcRo of all acquisition points at the current time point from the oil mist data set YW, and establishes the time evolution trend function YC; The time evolution trend function YC is obtained by the following formula: YC(t)={YcRo,1(t),YcRo,2(t),...,YcRo,n(t)}; Where YC(t) represents the time evolution trend function at time t, and n represents the total number of oil mist sampling points; The concentration trend value LC is obtained by constructing a concentration smoothing sequence using an exponential sliding filter; The concentration trend value LC is obtained by the following formula: LCi(t)=α*YcRo,i(t)+(1-α)*LCi(t-1); Where LCi(t) represents the concentration trend value of the i-th sampling point at time t, α represents the sliding filter coefficient, YcRo,i(t) represents the oil mist concentration characteristic of the i-th sampling point at time t, and LCi(t-1) represents the concentration trend value of the i-th sampling point at time t.

4. The intelligent purification system for centralized control of machine tool oil mist according to claim 3 is characterized in that: The purification deviation coefficient construction unit compares the obtained concentration trend value LC with the standard oil pollution threshold value Tc to calculate the purification deviation coefficient Hpx; The purification deviation coefficient Hpx is obtained by the following formula: Where, bc represents the risk magnification factor; By obtaining the purification deviation coefficient Hpx, the purification failure risk index R is calculated and the purification status of the oil collection point is judged by the purification failure risk index R; The purification failure risk index R is obtained by the following formula: In the formula, e represents a constant, and k represents the sensitivity factor that adjusts the steepness of the curve; The purification status is obtained by matching: When 0<purification failure risk index R<0.3, it indicates the first purification state and the purification is normal; When 0.3≤purification failure risk index R<0.6, it indicates the second purification state, fluctuations occur, and observation is required; When 0.6≤purification failure risk index R<1.0, it indicates the third purification state, purification is abnormal, and intervention is required.

5. The machine tool oil mist centralized control intelligent purification system according to claim 4 is characterized in that: The abnormal coupling identification module of the wind duct includes a local disturbance activation trigger unit and a non-steady-state coupling index calculation unit; The local disturbance activation trigger unit analyzes the purification failure risk index R. When a purification anomaly occurs, the collection point where the purification anomaly occurs is screened and marked, and the duct number corresponding to the collection point is used as the core area for wind pressure and wind speed analysis. The duct static pressure Pa and duct wind speed Sv in the core area are synchronously paired. The unsteady coupling index calculation unit differentiates and multiplies the paired duct static pressure Pa and duct wind speed Sv, and integrates them within a fixed period T to obtain the unsteady coupling coefficient FUx; The unsteady coupling coefficient FUx is obtained by the following formula: Where Pa(t) represents the static pressure of the duct at time t, Sv(t) represents the wind speed of the duct at time t, and d represents the derivative symbol; Compare the obtained non-steady-state coupling coefficient FUx with the preset abnormality identification threshold Tfu to determine the local disturbance state; The local disturbance state is obtained by matching: When the non-steady-state coupling coefficient FUx ≤ the abnormal identification threshold Tfu, it indicates that the local disturbance state is normal; When the non-steady-state coupling coefficient FUx is greater than the abnormal identification threshold Tfu, it indicates that a local disturbance has occurred and a recirculation zone exists.

6. The machine tool oil mist centralized control intelligent purification system according to claim 2 is characterized by: The filter load identification and attenuation assessment module includes a filter load state feature extraction unit and a performance attenuation index construction unit; The filter load state feature extraction unit takes the filter pressure difference Da and the oil mist volume Vu processed by the filter as the core input to construct the pressure difference growth factor Φ D and processing volume growth factor Φ V ; Pressure difference growth factor Φ D Obtained by the ratio of the filter pressure difference Da(t) at time t to the initial design pressure difference Do of the filter; Processing volume growth factor Φ V Obtained by the ratio of the oil mist volume Vu(t) processed by the filter at time t to the rated processing volume Vo of the filter; The pressure difference growth factor Φ obtained by the performance degradation index construction unit D and processing volume growth factor Φ V Perform fusion and calculate the filter performance attenuation index SZw; The filter performance degradation index SZw is obtained by the following formula: SZw=log(Φ D *Φ V +ES); Where, log represents the logarithmic function, ES represents a non-zero constant; The filter performance attenuation index SZw is obtained to identify the filter status and determine whether the purification capacity has decreased; The filter status is obtained by matching: When 0<filter performance attenuation index SZw<0.3, it indicates the first filter state, normal state; When 0.3≤filter performance attenuation index SZw<0.5, it indicates the second filter state, slight attenuation; When 0.5≤filter performance attenuation index SZw<0.8, it indicates the third filter state, critical state; When 0.8≤filter performance attenuation index SZw<1.0, it indicates the fourth filter state, overload state.

7. The machine tool oil mist centralized control intelligent purification system according to claim 6, characterized in that: The state control intensity calculation and strategy generation module includes a control intensity index calculation unit and a control strategy generation unit; The control intensity index calculation unit normalizes the obtained filter performance attenuation index SZw, non-steady-state coupling coefficient FUx, and purification deviation coefficient Hpx, eliminates the dimensions of different parameters, unifies them into the same dimension, obtains the normalized filter performance attenuation index gSZw, the normalized non-steady-state coupling coefficient gFUx, and the normalized purification deviation coefficient gHpx, and fuses them to obtain the comprehensive control intensity index ZHs; The comprehensive regulation intensity index ZHs is obtained by the following formula: ZHs=β1*gSZw+β2*gFUx+β3*gHpx; Where β1, β2, and β3 represent the preset weight values of the normalized filter performance attenuation index gSZw, the normalized non-steady-state coupling coefficient gFUx, and the normalized purification deviation coefficient gHpx, respectively, and β1+β2+β3≤1.

8. The machine tool oil mist centralized control intelligent purification system according to claim 7 is characterized by: The control strategy generation unit compares the obtained comprehensive control intensity index ZHs with the preset control threshold Tzh to obtain the control strategy AC; Control strategy AC is obtained by matching in the following ways: When the control intensity index ZHs is less than the control threshold Tzh*0.5, it indicates the normal zone, and the control strategy AC is: no action and small ventilation adjustment; When the control threshold Tzh*0.5≤control intensity index ZHs<control threshold Tzh, it indicates the steady-state control zone, and the control strategy AC: air volume fine-tuning and wind direction bias strategy; When the control threshold Tzh ≤ the control intensity index ZHs < the control threshold Tzh*1.5, it indicates the linkage control area, and the control strategy AC: filter replacement warning and path switching; When the control threshold Tzh*1.5≤control intensity index ZHs<control threshold Tzh*2.0, it indicates the warning response area, and the control strategy AC: deactivate some air sections, force filter replacement prompts and isolate abnormal points.

9. The machine tool oil mist centralized control intelligent purification system according to claim 8, characterized in that: The instruction execution and closed-loop state memory module includes a control instruction execution unit and an adaptive adjustment unit; The control instruction execution unit interprets the acquired control strategy AC as a device-level control instruction, driving the physical device to operate, including adjusting the air valve angle, switching the purification path, and lighting the filter indicator light. It also synchronously collects the effect information after the control execution through the feedback loop. After executing the control strategy AC, the adaptive adjustment unit conducts a trend assessment on the control execution results. When the purification deviation coefficient Hpx continues to decrease and the control intensity index ZHs is less than the control threshold Tzh, the dynamic downward adjustment mechanism of the standard oil pollution threshold Tc is triggered to obtain a new oil pollution threshold Tc. According to the purification deviation coefficient Hpx, the purification deviation change rate ΔHpx of N consecutive cycles is calculated; The purification deviation change rate ΔHpx is obtained by the following formula: Where N is the number of time window periods, Q is the summation index variable, Hpx(tQ) is the purification deviation coefficient at time tQ, and Hpx(t-Q+1) is the purification deviation coefficient at time t-Q+1. The new oil pollution threshold Tc is obtained by the following formula: Tc(t+1)=Tc(t)*(1-λ*ΔHpx); Where Tc(t+1) represents the standard oil contamination threshold at time t+1, Tc(t) represents the standard oil contamination threshold at time t, and λ represents the adjustment sensitivity coefficient.

10. A method for centralized control and intelligent purification of machine tool oil mist, applied to the centralized control and intelligent purification system for machine tool oil mist according to any one of claims 1 to 9, characterized in that: The following steps are involved: Step 1: The oil mist residual parameter acquisition and processing module collects oil mist data through the sensor installed at the end of the air duct, fits it into the initial data set CW, and performs preprocessing to obtain the oil mist data set YW; Step 2: The purification deviation identification and risk location module performs feature extraction on the oil mist dataset YW, extracts the current residual concentration, and compares it with the preset standard oil pollution threshold Tc to obtain the purification deviation coefficient Hpx and determine the purification status of the oil pollution collection point; Step 3: The duct abnormal coupling identification module analyzes the coordinated change trend of the duct static pressure Pa and the duct wind speed Sv according to the obtained purification status, obtains the non-steady-state coupling coefficient FUx, and identifies the local disturbance source and recirculation area; Step 4: The filter load identification and attenuation evaluation module analyzes the filter pressure difference Da and the oil mist volume Vu processed by the filter in the oil mist dataset YW to obtain the filter performance attenuation index SZw; Step 5: The state control intensity calculation and strategy generation module summarizes the obtained filter performance attenuation index SZw, non-steady-state coupling coefficient FUx and purification deviation coefficient Hpx to obtain the comprehensive control intensity index ZHs and output the control strategy AC; Step 6: The instruction execution and closed-loop state memory module performs specific control actions according to the acquired control strategy AC, and dynamically adjusts the standard oil pollution threshold Tc according to the control results.

Images