Methods, apparatus, equipment and storage media for maintaining the cleaning performance of cutting fluids.

Abstract

This invention relates to the field of cutting fluid cleaning performance maintenance technology, and in particular to a method, apparatus, device, and storage medium for cutting fluid cleaning performance maintenance. The method first acquires an initial multi-source physical dataset of the cutting fluid, performs data preprocessing on the initial multi-source physical dataset to obtain a target multi-source physical dataset, then calls a pre-trained cutting fluid performance judgment model to perform a safety performance judgment on the target multi-source physical dataset, and obtains a judgment result, which includes a safety judgment result and a warning level. Next, when the safety judgment result is a non-safety level, an execution instruction is generated based on the warning level. Finally, based on the execution instruction, an additive is injected into the cutting fluid to maintain its cleaning performance, aiming to replenish the consumed cleaning components and extend the service life of the cutting fluid.

Description

Technical Field

[0001] This invention relates to the field of cutting fluid cleaning performance maintenance technology, and in particular to a method, apparatus, equipment and storage medium for maintaining cutting fluid cleaning performance. Background Technology

[0002] In metal cutting and grinding processes, the cleaning performance of cutting fluids is crucial to production, directly affecting the removal of machining debris, oil, and other impurities. Insufficient cleaning performance leads to stubborn residues, which not only reduces workpiece quality and precision, accelerates wear on critical machine tool components, clogs related systems, and shortens equipment lifespan, but also causes frequent production interruptions, increases processing costs, poses safety hazards, and deteriorates the workshop environment. Existing technologies for improving the cleaning performance of cutting fluids and preventing impurity adhesion have significant limitations. On the one hand, optimizing the initial formula is static and cannot address the degradation of interfacial properties caused by the decay of effective components and the introduction of impurities during the use of cutting fluids. Various cutting fluids also require static compromises between lubrication, rust prevention, and cleaning properties, making it difficult to dynamically compensate for the individual decline in cleaning performance. On the other hand, commonly used maintenance methods such as concentration adjustment and complete fluid replacement are relatively crude. Adjusting the concentration sacrifices other properties of the cutting fluid, while complete fluid replacement is resource-intensive, costly, and can cause production interruptions. While current intelligent maintenance systems for cutting fluids can monitor common physicochemical parameters such as concentration and pH value and automatically compensate and add them to maintain the overall stability of the formula, the monitored parameters lack a direct and quantifiable correlation with the interface properties that determine cleaning performance. This makes it impossible to accurately diagnose and target the degradation of cleaning performance and to provide early warning of adhesion problems. Summary of the Invention

[0003] In order to overcome the shortcomings of the prior art, the present invention aims to provide a method, apparatus, equipment and storage medium for maintaining the cleaning performance of cutting fluid, which is intended to replenish the consumed cleaning components in a targeted manner and extend the service life of the cutting fluid.

[0004] The first aspect of this invention provides a method for maintaining the cleaning performance of a cutting fluid, comprising: acquiring an initial multi-source physical dataset of the cutting fluid; performing data preprocessing on the initial multi-source physical dataset to obtain a target multi-source physical dataset; calling a pre-trained cutting fluid performance judgment model to perform a safety performance judgment on the target multi-source physical dataset, obtaining a judgment result, the judgment result including a safety judgment result and a warning level; when the safety judgment result is a non-safety level, generating an execution instruction based on the warning level; and injecting an additive into the cutting fluid based on the execution instruction to maintain the cleaning performance of the cutting fluid.

[0005] Optionally, in a first implementation of the first aspect of the present invention, the initial multi-source physical dataset includes a subset of interface performance parameters; obtaining the initial multi-source physical dataset of the cutting fluid includes: measuring the surface tension of the cutting fluid using an industrial online sensor to obtain surface tension parameters; acquiring morphological images of the cutting fluid droplets in an observation window using a camera; performing contact angle calculation processing based on the morphological images using an image analysis tool to obtain contact angle parameters; and integrating the surface tension parameters and the contact angle parameters to obtain the subset of interface performance parameters.

[0006] Optionally, in a second implementation of the first aspect of the present invention, the initial multi-source physical dataset further includes an auxiliary parameter subset; after integrating the surface tension parameter and the contact angle parameter to obtain the interface performance parameter subset, the method further includes: measuring the conductivity of the cutting fluid using a conductivity sensor to obtain a conductivity parameter; measuring the acidity / alkalinity of the cutting fluid using a pH sensor to obtain an acidity / alkalinity parameter; measuring the liquid turbidity of the cutting fluid using a turbidity sensor to obtain a liquid turbidity parameter; measuring the refractive index of the cutting fluid using a refractometer to obtain a refractive index reading parameter; and integrating the conductivity parameter, the acidity / alkalinity parameter, the liquid turbidity parameter, and the refractive index reading parameter to obtain the auxiliary parameter subset.

[0007] Optionally, in a third implementation of the first aspect of the present invention, the step of preprocessing the initial multi-source physical dataset to obtain a target multi-source physical dataset includes: performing unit unification processing on the initial multi-source physical dataset using a physical quantity conversion algorithm to obtain a unified physical dataset; performing filtering processing on the unified physical dataset using a moving average algorithm to obtain a filtered physical dataset; and performing standardization processing on the filtered physical dataset using a standardization algorithm to obtain the target multi-source physical dataset.

[0008] Optionally, in a fourth implementation of the first aspect of the present invention, the cutting fluid performance judgment model includes a first threshold comparison module, a second threshold comparison module, a performance safety judgment module, and a warning level evaluation module, wherein the first threshold comparison module, the second threshold comparison module, the performance safety judgment module, and the warning level evaluation module are connected sequentially; the step of calling the pre-trained cutting fluid performance judgment model to perform a safety performance judgment on the target multi-source physical dataset and obtaining a judgment result, wherein the judgment result includes a safety judgment result and a warning level, includes: obtaining a surface tension threshold stored in the first threshold comparison module, comparing the surface tension threshold with the surface tension parameter based on the first threshold comparison module to obtain a first comparison result; obtaining a contact angle threshold stored in the second threshold comparison module, comparing the contact angle threshold with the contact angle parameter based on the second threshold comparison module to obtain a second comparison result; performing a performance safety judgment on the cutting fluid based on the first comparison result and the second comparison result using the performance safety judgment module to obtain the safety judgment result; when the safety judgment result is a non-safety level, performing a warning level evaluation on the safety performance of the cutting fluid based on the first comparison result and the second comparison result using the warning level evaluation module to obtain the warning level.

[0009] Optionally, in a fifth implementation of the first aspect of the present invention, when the safety determination result is a non-safety level, the step of evaluating the safety performance of the cutting fluid based on the first comparison result and the second comparison result by the warning level evaluation module to obtain the warning level includes: when the safety determination result is a non-safety level, obtaining the number of parameters exceeding the threshold, the surface tension deviation threshold parameter, and the contact angle deviation threshold parameter based on the first comparison result and the second comparison result respectively; and evaluating the safety performance of the cutting fluid based on the number of parameters exceeding the threshold, the surface tension deviation threshold parameter, and the contact angle deviation threshold parameter by the warning level evaluation module to obtain the warning level.

[0010] Optionally, in a sixth implementation of the first aspect of the present invention, the step of generating an execution instruction based on the warning level when the safety determination result is a non-safety level includes: when the safety determination result is a non-safety level, using an additive selection algorithm to perform additive selection processing according to the warning level to obtain a target additive type; using a quantitative adjustment algorithm to calculate the additive dosage according to the warning level and the target additive type to obtain a compensation dosage corresponding to the target additive type; and generating the execution instruction based on the target additive type and the compensation dosage.

[0011] A second aspect of the present invention provides a cutting fluid cleaning performance maintenance device, comprising: a parameter monitoring module for acquiring an initial multi-source physical dataset of the cutting fluid; a data preprocessing module for preprocessing the initial multi-source physical dataset to obtain a target multi-source physical dataset; a safety performance judgment module for calling a pre-trained cutting fluid performance judgment model to perform a safety performance judgment on the target multi-source physical dataset and obtain a judgment result, the judgment result including a safety judgment result and a warning level; an instruction generation module for generating an execution instruction based on the warning level when the safety judgment result is a non-safety level; and an instruction execution module for injecting an additive into the cutting fluid based on the execution instruction to maintain the cleaning performance of the cutting fluid.

[0012] A third aspect of the present invention provides a cutting fluid cleaning performance maintenance device, the cutting fluid cleaning performance maintenance device comprising: a memory and at least one processor, the memory storing instructions; the at least one processor calling the instructions in the memory to cause the cutting fluid cleaning performance maintenance device to perform each step of the cutting fluid cleaning performance maintenance method described in any of the preceding claims.

[0013] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the steps of the cutting fluid cleaning performance maintenance method described in any of the preceding claims.

[0014] In the technical solution of this invention, firstly, an initial multi-source physical dataset of the cutting fluid is obtained. The initial multi-source physical dataset is preprocessed to obtain a target multi-source physical dataset. Then, a pre-trained cutting fluid performance judgment model is called to judge the safety performance of the target multi-source physical dataset, and the judgment result is obtained. The judgment result includes a safety judgment result and a warning level. Next, when the safety judgment result is a non-safety level, an execution instruction is generated based on the warning level. Finally, based on the execution instruction, an additive is injected into the cutting fluid to maintain the cleaning performance of the cutting fluid, aiming to replenish the consumed cleaning components and extend the service life of the cutting fluid. Attached Figure Description

[0015] Figure 1 A logic flowchart of the cutting fluid cleaning performance maintenance method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the cutting fluid cleaning performance maintenance device provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the cutting fluid cleaning performance maintenance device provided in an embodiment of the present invention. Detailed Implementation

[0016] This invention provides a method, apparatus, device, and storage medium for maintaining the cleaning performance of cutting fluid. In this invention, the terms "first," "second," "third," "fourth," etc. (if applicable) in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0017] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 One embodiment of the cutting fluid cleaning performance maintenance method in this invention includes: 101. Obtain the initial multi-source physical dataset of the cutting fluid; In this embodiment, an initial multi-source physical dataset of the cutting fluid is obtained to monitor the interfacial physical parameters of the cutting fluid. Related auxiliary physicochemical parameters can be selected and integrated according to actual application needs. The system collects and integrates these parameters to form a multi-dimensional initial multi-source physical dataset. The monitoring of interfacial physical parameters includes the quantitative measurement of the cutting fluid's surface tension and contact angle. Surface tension is a direct physical measure of the cutting fluid's surface contraction ability; its value is negatively correlated with the cutting fluid's wetting, spreading, and penetrating abilities, making it a key quantitative indicator characterizing the basic interfacial performance of the cutting fluid. The contact angle is obtained by measuring the cutting fluid on solid surfaces of simulated machine tool inner walls, such as cast iron and cast steel. This parameter directly reflects the cutting fluid's wetting ability on solid surfaces; its value is negatively correlated with wettability, directly reflecting the cutting fluid's ability to remove dirt from solid surfaces and its core characteristic of resisting its own residue. Simultaneously, it can integrate optical characteristic parameters such as conductivity, pH value, turbidity, and refractive index as needed. Although these parameters cannot directly characterize the interfacial performance of the cutting fluid, their numerical changes can effectively indicate intrinsic state changes within the cutting fluid system, such as changes in composition, introduction of contaminants, and performance degradation. This provides multi-dimensional physicochemical information support for the comprehensive diagnosis of cutting fluid cleaning performance. The initial multi-source physical dataset comprehensively captures the physicochemical characteristics of the cutting fluid from both the dual dimensions of interfacial performance characterization and overall system state indication. This achieves accurate quantification of core indicators related to cutting fluid cleaning performance and enhances the comprehensiveness and reliability of data acquisition by relying on the complementarity of multi-source parameters. It effectively avoids the one-sidedness of single-parameter monitoring and lays a solid and comprehensive physical data foundation for the subsequent accurate determination of cutting fluid cleaning performance, scientific evaluation of early warning levels, and formulation of targeted intervention strategies. This ensures the accuracy and scientific nature of subsequent full-process data analysis and performance judgment.

[0018] 102. Perform data preprocessing on the initial multi-source physical dataset to obtain the target multi-source physical dataset; In this embodiment, a systematic data preprocessing is performed on the initial multi-source physical dataset. Through sequential unit unification, filtering, and standardization operations, the original data is optimized and its quality improved, ultimately yielding the target multi-source physical dataset. Specifically, unit unification addresses the differences in the original measurement forms of different parameters in the initial multi-source physical dataset, completing the standardization and conversion of the measurement units for all parameters, ensuring that all physical parameters are presented in standard physical dimensions, and eliminating data analysis bias caused by inconsistent measurement forms. Filtering identifies and removes abnormal jumps in data caused by equipment operation fluctuations and on-site environmental interference during data acquisition, ensuring the continuity, stability, and authenticity of the data sequence. Standardization then applies dimensionless processing to the dataset after unit unification and filtering, eliminating data analysis weight bias caused by significant differences in dimensional attributes and numerical ranges between different physical parameters, ensuring that each parameter has equal data analysis comparability. The target multi-source physical dataset obtained through the above preprocessing effectively filters out invalid interference information in the original data, realizes the standardization of data format and the unification of analysis dimensions, greatly improves the overall quality and effective utilization of data, avoids the problem of distortion of subsequent cutting fluid performance judgment results due to differences in the format and abnormal deviations of the original data from the source of data, and can truly, accurately and comprehensively reflect the actual physicochemical state of the cutting fluid.

[0019] 103. Call the pre-trained cutting fluid performance judgment model to perform safety performance judgment on the target multi-source physical dataset, and obtain the judgment result, which includes the safety judgment result and the warning level; In this embodiment, the cutting fluid performance judgment model of the present invention is a hybrid judgment model that integrates expert rules with lightweight machine learning, with expert rules as the core framework to ensure interpretability. Model training is conducted in stages based on experimental and field data: first, experimental calibration data and historical industrial field data for different cutting fluid types and operating conditions are collected, and after preprocessing, the datasets are divided and labeled with level labels. A two-layer framework is built: the expert rule layer solidifies the dual-parameter threshold comparison and safety judgment logic, and the machine learning layer constructs an early warning classification network. Cold start training is completed using experimental data, and then the thresholds and algorithm weights are fine-tuned using field data. After testing and verification, it is deployed. Subsequently, online incremental training is conducted based on new data recorded in real time by the system, dynamically optimizing the thresholds and model parameters to achieve continuous performance improvement with data accumulation, balancing judgment rigor and operating condition adaptability.

[0020] In this embodiment, a pre-trained cutting fluid performance assessment model is invoked to perform safety performance assessments on a target multi-source physical dataset. The final result includes both a safety assessment and a warning level. The cutting fluid performance assessment model pre-sets upper limits for surface tension and contact angle for different cutting fluid types (fully synthetic, semi-synthetic, etc.) and specific operating conditions determined by the customer based on the main processing materials. These thresholds are set based on historical experimental data and relevant industry standards, and authorized personnel can adaptively fine-tune the thresholds within a specified range. The model simultaneously stores various threshold parameters and records historical data. Furthermore, the cutting fluid performance assessment model incorporates a historical database. This database continuously collects and retains full monitoring parameters of the cutting fluid, detailed information on various maintenance events, and performance change curves after maintenance. Through continuous accumulation and analysis of this data, the model mines and extracts the performance degradation patterns of the cutting fluid, providing data support for the model's comprehensive assessment. During model execution, the preset thresholds that are compatible with the current cutting fluid type and actual working conditions are first retrieved from the threshold model. The surface tension and contact angle core interface parameters in the target multi-source physical dataset are quantitatively compared and analyzed with the corresponding thresholds. Then, combined with the performance degradation law of cutting fluid under similar working conditions stored in the historical database, the implementation effect of past maintenance operations, and performance change characteristics, the actual cleaning performance status of the cutting fluid is comprehensively judged from multiple dimensions to obtain the corresponding safety judgment result. At the same time, based on the degree of deviation of the core parameters from the thresholds and combined with the synergistic change characteristics of multi-source auxiliary parameters, the risk level of cutting fluid performance degradation is graded and assessed, and finally the corresponding warning level is determined. The support of historical databases allows the model to fully combine the actual performance change patterns of cutting fluid with on-site maintenance experience, making the safety judgment results and warning levels more scientific and accurate. The fine-tunable thresholds improve the model's scenario adaptability and operational flexibility. The continuous accumulation of historical data also provides real and effective industrial data basis for the model's subsequent iterative optimization. It can accurately identify and grade warnings of the deterioration of cutting fluid cleaning performance, capture abnormal performance trends in a timely manner, and provide a scientific basis for subsequent targeted maintenance interventions, effectively avoiding the problems of untimely or excessive maintenance caused by performance judgment deviations.

[0021] 104. When the security determination result is a non-security level, an execution instruction is generated based on the warning level; In this embodiment, when the safety assessment result is unsafe, the system will rely on the clearly defined risk level of the warning level, combined with the warning level, actual processing conditions, and historical maintenance data, to determine the target additive type suitable for cleaning performance restoration. This ensures the compatibility of the additive with the cutting fluid system. Simultaneously, based on the specific degree to which the interface parameters deviate from the preset threshold in the warning level, the system will accurately calculate the additive compensation dosage to ensure a high degree of matching between the dosage and the performance degradation level. This satisfies the performance restoration requirements while avoiding system imbalance caused by excessive addition. Furthermore, the system only targets the degradation of cleaning performance, without affecting the original core performance of the cutting fluid, such as lubrication and rust prevention. Based on this, the system integrates the target additive type and the calculated compensation dosage to generate standardized execution instructions, clarifying various operating parameters. This provides a clear and precise action basis for the subsequent metering injection unit, achieving refined control of additive dosage and avoiding performance restoration failures or system contamination problems caused by inaccurate dosage in traditional extensive maintenance.

[0022] 105. Based on the execution command, inject an additive into the cutting fluid to maintain the cleaning performance of the cutting fluid.

[0023] In this embodiment, after the system responds to the execution command, the additive storage tank stably supplies liquid to the precision metering and injection unit. The storage tank is made of chemically corrosion-resistant material and can safely store high-concentration functional additive concentrate. Its capacity is scientifically designed according to the on-site maintenance frequency to meet the continuous use needs for several weeks to several months. It is also equipped with a liquid level sensor to monitor the liquid level in real time and trigger a low liquid level alarm to ensure continuous liquid supply and avoid maintenance interruptions. The storage tank is connected to the precision metering and injection unit through a dedicated pipeline to ensure smooth additive delivery. The precision metering and injection unit, as the core execution component, uses a high-precision metering pump as its core. Its flow rate is precisely controlled according to the execution command to achieve accurate metering of the additive compensation dosage, avoiding imbalance in the cutting fluid system caused by dosage deviation. The metered additive is transported to the designated injection location through pipelines, such as the Venturi tube injection point connected to the main circulation pipeline of the cutting fluid, or directly introduced into the turbulent zone of the liquid tank. The fluid dynamics effect promotes rapid and thorough homogeneous mixing of the additive and the cutting fluid, ensuring the uniformity of the repair effect. Each action is recorded electronically, preserving information such as the time of addition, dosage, and equipment operating status, thus enabling full traceability of the maintenance process.

[0024] The effectiveness of the cutting fluid cleaning performance maintenance method of this invention has been fully supported by multiple sets of experiments. First, screening and effect verification experiments of functional additives were conducted. Using a semi-synthetic cutting fluid working fluid that had been used continuously for two weeks in a machining workshop and showed a slight tendency to adhere as the base fluid, fatty alcohol polyoxyethylene ether, alkylphenol polyoxyethylene ether, glycerol, and a blank control were selected as candidate additives. After adding 0.1 wt% of each candidate substance under the same conditions, the surface tension, contact angle, and foam volume of the samples were measured. Experimental results showed that fatty alcohol polyoxyethylene ether significantly reduced the surface tension and contact angle of the cutting fluid, improved the wettability and spreadability of the liquid on the wall surface, and its additional properties such as emulsification, dispersion, and low adsorption further avoided the problems of local adhesion and residual drying, ultimately achieving effective improvement in wall adhesion. Its interface performance repair effect was far superior to other comparative samples, and the foam volume was within a controllable range. Although glycerol had some improvement effect, the effect was limited. Therefore, fatty alcohol polyoxyethylene ether was determined to be the preferred functional additive for this system.

[0025] Based on this, a correlation verification experiment was conducted between the interface performance threshold and the adhesion risk. By adding different proportions of simulated consumables or surfactants to fresh semi-synthetic cutting fluid, samples with varying surface tension and contact angle gradients were prepared. The adhesion tendency of each sample was quantified using the dirt suspension adhesion method. The results showed that when the surface tension was below approximately 40 mN / m and the contact angle was below approximately 45°, the amount of dirt adhesion remained at a low level. Once this critical point was exceeded, the amount of adhesion increased sharply. Based on this, the warning threshold for semi-synthetic cutting fluid was set at an upper limit of 40 mN / m for surface tension and an upper limit of 45° for contact angle, providing a scientific basis for the system's warning mechanism.

[0026] Finally, a comparative verification experiment was conducted between the method of this invention and the traditional manual maintenance method. In two parallel single-machine circulation systems of cutting fluid, the experimental group was connected to the monitoring and addition prototype system of this invention. The aforementioned warning threshold was set, and 0.1 wt% fatty alcohol polyoxyethylene ether concentrate was automatically added when the parameters exceeded the standard. The control group used the traditional method of manually checking the concentration every 24 hours and only adding diluent. After 240 hours of continuous operation, the amount of dirt adhering in the experimental group was reduced by approximately 78% compared to the control group, and the interfacial performance of the cutting fluid remained at a good level throughout the entire cycle. The additive consumption was only 0.4 kg, while the surface tension and contact angle of the control group increased significantly, and the amount of dirt adhering increased significantly. The experimental results fully verify the significant advantages of the cutting fluid cleaning performance maintenance method provided by this invention in terms of anti-adhesion effect, performance maintenance efficiency, and economic cost control. Its precise on-demand addition mode can effectively prevent the formation of an adhesive layer and avoid system imbalance caused by over-addition. The fully traceable electronic records also provide data support for the optimization of subsequent maintenance strategies.

[0027] In this embodiment of the invention, the initial multi-source physical dataset includes a subset of interface performance parameters; the acquisition of the initial multi-source physical dataset of the cutting fluid includes: measuring the surface tension of the cutting fluid using an industrial online sensor to obtain surface tension parameters; acquiring morphological images of the cutting fluid droplets in an observation window using a camera; performing contact angle calculation processing based on the morphological images using an image analysis tool to obtain contact angle parameters; and integrating the surface tension parameters and the contact angle parameters to obtain the subset of interface performance parameters.

[0028] In this embodiment, an initial multi-source physical dataset of the cutting fluid is acquired. Online monitoring is the preferred method for acquiring the subset of interface performance parameters, supplemented by offline monitoring as a means of verification and calibration, balancing the real-time nature and accuracy of data acquisition. Specifically, surface tension parameters are acquired online using an industrial online sensor. This sensor is designed based on the maximum bubble pressure method, calculating surface tension by measuring the maximum pressure required to form a bubble at the capillary end immersed in the cutting fluid. It boasts advantages such as strong anti-contamination capability and high stability, effectively avoiding the influence of complex industrial fluid environments containing impurities such as oil and particulate matter, and is virtually unaffected by liquid color and turbidity. The sensor is adaptable to installation in the main or bypass pipeline of the cutting fluid, enabling continuous or intermittent automatic measurement and accurately capturing quantitative data reflecting the surface shrinkage capacity of the cutting fluid. Offline mode serves as a supplement and calibration basis for the online solution. Samples are periodically taken from the fluid tank, and precise measurements are performed using a laboratory surface tension meter via the platinum plate method or platinum ring method. This method provides high data accuracy, serving both as a calibration tool for the online sensor and as reliable data support for establishing initial benchmark values ​​for the system.

[0029] In this embodiment, the acquisition of contact angle parameters also adopts a collaborative online and offline mode. The online solution relies on optical imaging technology. A monitoring structure with a standard material observation window is set in the circulation system. The observation window is made of cast iron, cast steel, or other materials consistent with the inner wall of the machine tool to simulate the liquid-solid contact environment under actual working conditions. At the same time, an automatic cleaning and dripping device is provided to prevent the window from being contaminated by cutting fluid impurities to ensure measurement accuracy. A high-definition camera automatically captures the morphological image of cutting fluid droplets on the clean observation window surface. Then, a professional image analysis tool (such as ImageJ or the Liquid Bridge-Axisymmetric Droplet Shape Analysis Plugin) processes the morphological image, analyzes the droplet contour features in real time, and calculates the contact angle parameter, which intuitively reflects the wetting ability of the cutting fluid on the metal surface. The offline practical solution involves periodic sampling and measurement using a laboratory contact angle measuring instrument on a standardized metal sample with the same material as the inner wall of the machine tool. This can be used as a calibration benchmark for the online monitoring solution or as the main monitoring method to be performed periodically when online monitoring conditions are not available.

[0030] In this embodiment, the surface tension and contact angle parameters obtained from online and offline measurements are integrated, and abnormal data is removed to form a complete subset of interface performance parameters. These parameters comprehensively characterize the cleaning, spreading, and penetrating capabilities of the cutting fluid from two core dimensions: the liquid's own cohesive force and the liquid-solid interface wetting characteristics, achieving multi-dimensional and accurate quantification of interface performance. This acquisition method, through the coordinated use of online and offline monitoring, ensures the efficiency and real-time nature of data acquisition in continuous industrial production scenarios through the online mode, while leveraging the high precision of the offline mode to calibrate data and establish benchmarks. This effectively compensates for the limitations of a single monitoring mode and avoids problems such as online sensor drift and offline measurement lag. The resulting subset of interface performance parameters more realistically and comprehensively reflects the actual interface state of the cutting fluid.

[0031] In this embodiment of the invention, the initial multi-source physical dataset further includes an auxiliary parameter subset; after integrating the surface tension parameter and the contact angle parameter to obtain the interface performance parameter subset, the method further includes: measuring the conductivity of the cutting fluid using a conductivity sensor to obtain a conductivity parameter; measuring the acidity / alkalinity of the cutting fluid using a pH sensor to obtain an acidity / alkalinity parameter; measuring the liquid turbidity of the cutting fluid using a turbidity sensor to obtain a liquid turbidity parameter; measuring the refractive index of the cutting fluid using a refractometer to obtain a refractive index reading parameter; and integrating the conductivity parameter, the acidity / alkalinity parameter, the liquid turbidity parameter, and the refractive index reading parameter to obtain the auxiliary parameter subset.

[0032] In this embodiment, the initial multi-source physical dataset also includes an auxiliary parameter subset. This subset supplements information from the overall chemical state dimension of the cutting fluid system, forming a multi-dimensional data support system with the interface performance parameter subset. After integrating the surface tension parameter and the contact angle parameter to obtain the interface performance parameter subset, the auxiliary parameters are further collected and integrated. A conductivity sensor is used to continuously measure the conductivity of the cutting fluid. The obtained conductivity parameter reflects the total ion concentration in the solution. Its continuous increase usually originates from the accumulation of mineral salts, which come from water evaporation, workpiece introduction, additive decomposition, or hard water reaction products. Excessively high ionic strength can interfere with the effectiveness of corrosion inhibitors, affect the micelle morphology of some surfactants, and indirectly negatively impact cleaning and rust prevention performance. A pH sensor is used to measure the acidity and alkalinity of the cutting fluid in real time. The obtained acidity and alkalinity parameter is a core indicator of the system's chemical equilibrium and biological stability. A decrease in pH value, i.e., acidification, is usually accompanied by active microbial growth, which can cause the cutting fluid to spoil and smell bad, its rust prevention performance to drop sharply, some additives to become ineffective, and even exacerbate metal corrosion. Turbidity sensors are used to measure the turbidity of cutting fluids. The obtained turbidity parameters visually reflect the content of insoluble suspended particles using optical methods. Abnormally high turbidity indicates an increase in particles such as metal powder, putrefactive products, or soap formation, which is a direct signal of decreased cleanliness and deposition risk. Refractometers are used to measure the refractive index of the cutting fluid. The obtained refractive index readings are commonly used for rapid on-site estimation of the concentration of effective components. Abnormally low readings may indicate insufficient concentration or a large amount of contaminating oil, while high or unclear readings suggest excessively high impurity concentration, oil-water separation, or severe instability of the emulsion. Contaminating oil itself is also a major contaminant that causes system viscosity and promotes adhesion. Integrating the above conductivity, pH, turbidity, and refractive index reading parameters forms a complete subset of auxiliary parameters. Changes in the values ​​of these auxiliary parameters can effectively indicate the internal state of changes in system composition, introduction of contaminants, or performance degradation. Working synergistically with the subset of interface performance parameters, this significantly improves the comprehensiveness and accuracy of cutting fluid performance diagnosis. For example, abnormal fluctuations in conductivity and pH can provide early warnings of microbial contamination or salt accumulation risks, offering a precursor signal for interfacial performance degradation. Changes in turbidity and refractive index readings can directly reflect the cleanliness and concentration stability of the system, helping to verify whether abnormalities in interfacial performance parameters are caused by contamination or concentration imbalance. Through complementary verification using multi-source data, the system can more accurately pinpoint the root cause of performance degradation, avoiding the limitations and misjudgment risks associated with monitoring a single parameter.

[0033] In this embodiment of the invention, the step of preprocessing the initial multi-source physical dataset to obtain the target multi-source physical dataset includes: using a physical quantity conversion algorithm to perform unit unification processing on the initial multi-source physical dataset to obtain a unified physical dataset; using a moving average algorithm to perform filtering processing on the unified physical dataset to obtain a filtered physical dataset; and using a standardization algorithm to perform standardization processing on the filtered physical dataset to obtain the target multi-source physical dataset.

[0034] In this embodiment, a physical quantity conversion algorithm (such as a linear conversion algorithm) is first used to unify the units, converting the original measurement forms of different types of parameters in the initial dataset into standard physical dimensions. For example, the original pixel values ​​obtained from image analysis in contact angle monitoring are converted into precise angle values ​​through the algorithm. At the same time, the measurement units of interface performance parameters and auxiliary parameters are standardized to eliminate measurement deviations caused by differences in the original data acquisition forms, ensuring that all data are presented in a standardized physical quantity form, laying the foundation for subsequent unified analysis.

[0035] In this embodiment, after unifying the units to obtain a unified physical dataset, a moving average algorithm (such as a moving average algorithm) is used for filtering. For continuously collected parameter data, especially core parameters such as surface tension that are susceptible to equipment operation fluctuations and environmental interference, the moving average algorithm smooths the data sequence, effectively eliminating abnormal jumps in data generated during the collection process, weakening the impact of random interference on data accuracy, and extracting stable trend values ​​of data changes. This makes the processed data more reflective of the true state of the cutting fluid performance and avoids abnormal data misleading subsequent analysis and decision-making.

[0036] In this embodiment, a standardization algorithm (such as the Z-score standardization algorithm) is finally used to process the filtered physical dataset to eliminate the bias in analysis weights caused by the large differences in the dimensional properties and numerical ranges of different parameters. For example, the dimensional and numerical ranges of parameters such as surface tension, conductivity, and turbidity are quite different. After standardization, each parameter can have the same data analysis comparability, ensuring that the subsequent performance judgment model can conduct a fair and accurate comprehensive evaluation of multi-dimensional parameters.

[0037] In this embodiment of the invention, the cutting fluid performance judgment model includes a first threshold comparison module, a second threshold comparison module, a performance safety judgment module, and a warning level evaluation module, which are sequentially connected. The pre-trained cutting fluid performance judgment model is invoked to perform a safety performance judgment on the target multi-source physical dataset, obtaining a judgment result. The judgment result includes a safety judgment result and a warning level, comprising: acquiring a surface tension threshold stored in the first threshold comparison module; comparing the surface tension threshold with a surface tension parameter based on the first threshold comparison module to obtain a first comparison result; acquiring a contact angle threshold stored in the second threshold comparison module; comparing the contact angle threshold with a contact angle parameter based on the second threshold comparison module to obtain a second comparison result; performing a performance safety judgment on the cutting fluid based on the first and second comparison results using the performance safety judgment module to obtain a safety judgment result; and when the safety judgment result is a non-safety level, performing a warning level evaluation on the safety performance of the cutting fluid based on the first and second comparison results using the warning level evaluation module to obtain the warning level.

[0038] In this embodiment, the cutting fluid performance judgment model completes the safety performance judgment of the target multi-source physical dataset through the sequential linkage of the first threshold comparison module, the second threshold comparison module, the performance safety judgment module, and the early warning level evaluation module. The final output includes both the safety judgment result and the early warning level. This invention's cutting fluid performance judgment model is a hybrid judgment model that integrates expert rules with lightweight machine learning, using expert rules as the core framework to ensure interpretability. Model training is conducted in stages based on experimental and field data: first, experimental calibration data and historical industrial field data for different cutting fluid types and operating conditions are collected, preprocessed, and then the dataset is divided and labeled with level labels. A two-layer framework is built: the expert rule layer solidifies the dual-parameter threshold comparison and safety judgment logic, and the machine learning layer constructs the early warning level network. Cold start training is completed using experimental data, and then the thresholds and algorithm weights are fine-tuned using field data. After testing and verification, it is deployed. Subsequently, online incremental training is performed based on new data recorded in real time by the system, dynamically optimizing the thresholds and model parameters to achieve continuous performance improvement with data accumulation, balancing judgment rigor and operating condition adaptability.

[0039] During model execution, the first threshold comparison module retrieves a preset surface tension threshold. This threshold is designed for different cutting fluid types, such as fully synthetic and semi-synthetic, taking into account the specific working conditions determined by the customer's main processing materials. It is based on prior experimental data and industry standards, and allows authorized personnel to make adaptive fine-tuning within a reasonable range. For example, the preset upper limit of surface tension threshold is 40 mN / m for semi-synthetic cutting fluid and 35 mN / m for fully synthetic cutting fluid. The first threshold comparison module quantitatively compares the retrieved customized surface tension threshold with the surface tension parameters in the target multi-source physical dataset, outputting the first comparison result to determine whether the surface tension parameters exceed the threshold and the degree of deviation.

[0040] Subsequently, the second threshold comparison module synchronously retrieves the contact angle threshold under the corresponding working condition. This threshold also follows a customized setting logic, combining the cutting fluid type and machining condition presets. For example, the upper limit threshold for the contact angle of semi-synthetic cutting fluid is set to 45°, and that of fully synthetic cutting fluid is set to 35°, and can be fine-tuned and calibrated by authorized personnel. The second threshold comparison module compares and analyzes the contact angle threshold with the contact angle parameters in the target dataset to obtain a second comparison result, accurately reflecting the threshold deviation of the contact angle parameters.

[0041] The performance and safety assessment module makes a comprehensive judgment based on the above two sets of comparison results. When the surface tension and contact angle parameters do not exceed the corresponding thresholds, the cutting fluid interface performance is judged to be good, and the safety assessment result of the safety level is output. When any parameter exceeds the threshold or both parameters exceed the threshold, it is judged to be unsafe, indicating that the cutting fluid has problems with reduced cleaning performance and increased adhesion risk.

[0042] When the safety assessment result is unsafe, the warning level assessment module further combines the two sets of comparison results and the degree of parameter deviation from the threshold to conduct a graded assessment. Simultaneously, it references monitoring parameters, maintenance events, and post-maintenance performance change curves stored in the historical operation database under similar operating conditions. Based on the tracked performance degradation patterns, the assessment accuracy is optimized. If only a single parameter exceeds the threshold and the deviation is small, it is assessed as a medium-level warning; if both parameters exceed the threshold simultaneously or a single parameter deviates significantly from the threshold, it is assessed as a high-level alarm, and the warning level is ultimately output. This approach ensures the comprehensiveness of performance safety assessment through dual-parameter collaborative comparison and achieves precise warning level classification based on the degree of deviation and historical data. This provides a clear basis for subsequent targeted maintenance interventions. Furthermore, the continuous accumulation of historical data can feed back into the iterative optimization of the threshold model, forming a data-driven virtuous cycle. This effectively avoids problems such as untimely or excessive maintenance due to judgment errors, ensuring the accuracy and efficiency of cutting fluid performance control.

[0043] In this embodiment of the invention, when the safety determination result is a non-safety level, the step of evaluating the safety performance of the cutting fluid based on the first comparison result and the second comparison result by the warning level evaluation module to obtain the warning level includes: when the safety determination result is a non-safety level, obtaining the number of parameters exceeding the threshold, the surface tension deviation threshold parameter, and the contact angle deviation threshold parameter based on the first comparison result and the second comparison result respectively; and evaluating the safety performance of the cutting fluid based on the number of parameters exceeding the threshold, the surface tension deviation threshold parameter, and the contact angle deviation threshold parameter by the warning level evaluation module to obtain the warning level.

[0044] In this embodiment, when the safety assessment result is non-safe, the early warning level assessment module first extracts core assessment indicators from the first comparison result and the second comparison result, clarifying the number of parameters exceeding the threshold, the surface tension deviation from the threshold parameter, and the contact angle deviation from the threshold parameter. The number of parameters exceeding the threshold is the number of items in the two core parameters, surface tension and contact angle, that exceed the corresponding preset threshold. The surface tension and contact angle deviation from the threshold parameters quantify the difference between the actual values ​​of the two parameters and the customized thresholds, accurately representing the degree of parameter deviation and providing a quantitative basis for risk classification. The early warning level assessment module conducts a comprehensive assessment based on the above three indicators, and combines the adaptation threshold characteristics of the cutting fluid type and processing conditions to construct a quantitative classification logic: if only a single parameter exceeds the threshold, and the corresponding deviation from the threshold parameter is within a small range, it is assessed as a medium-level early warning, indicating that the cutting fluid has a slight decrease in cleaning performance and a low risk of adhesion; if both parameters exceed the threshold simultaneously, or if a single parameter deviates from the threshold parameter to a significant level, it is assessed as a high-level alarm, indicating that the cutting fluid's cleaning performance is severely insufficient, the risk of adhesion is high, and immediate intervention is required. This assessment method uses quantitative indicators as its core, abandons the extensive grading logic, and achieves accurate definition of the warning level through the synergistic analysis of the number of thresholds exceeded and the degree of deviation. This avoids the one-sidedness of single indicator assessment and can accurately match the actual performance degradation state.

[0045] In this embodiment of the invention, the step of generating an execution instruction based on the warning level when the safety determination result is a non-safety level includes: when the safety determination result is a non-safety level, using an additive selection algorithm to perform additive selection processing according to the warning level to obtain a target additive type; using a quantitative adjustment algorithm to calculate the additive dosage according to the warning level and the target additive type to obtain a compensation dosage corresponding to the target additive type; and generating the execution instruction based on the target additive type and the compensation dosage.

[0046] In this embodiment, the additive selection algorithm (such as the CBR algorithm, i.e., Case-Based Reasoning Algorithm) relies on the historical database built into the model and the results of previous experimental verification. It matches the appropriate target additives according to the warning level, giving priority to functional additives that have been proven in practice to effectively repair interface performance and have excellent compatibility with the corresponding cutting fluid system. At the same time, it also takes into account the degradation characteristics corresponding to the warning level, ensuring that the selected additives can accurately target and improve the problem of insufficient cleaning performance, avoid adverse reactions with the original components of the cutting fluid, and maintain the overall stability of the system.

[0047] In this embodiment, the quantitative adjustment algorithm (such as the proportional-integral algorithm) uses the warning level as the core basis, and simultaneously combines parameters such as the active concentration and efficiency of the target additive type and the current volume of the cutting fluid, and calls the preset quantitative adjustment algorithm to accurately calculate the compensation dosage. The algorithm quantifies the degree of deviation of surface tension and contact angle from the threshold, and combines it with the risk level corresponding to the warning level. A medium warning corresponds to a slight deviation of a single parameter, and the basic compensation dosage is calculated; a high warning corresponds to two parameters exceeding the threshold or a single parameter deviating significantly, and the increment is dynamically adjusted according to the degree of deviation on the basic dosage to ensure that the dosage is highly matched with the degree of performance degradation, which not only meets the cleaning performance repair needs, but also avoids secondary problems such as abnormal foaming and interference with lubrication performance caused by excessive addition. Based on the determined target additive type and the accurately calculated compensation dosage, a standardized execution instruction is integrated to clarify each core operation parameter, and is sent synchronously to the execution addition module to provide a clear and reliable action basis for subsequent metering and injection. Through algorithmic selection and quantitative calculation, refined control of maintenance intervention is achieved, abandoning the blindness of traditional extensive addition.

[0048] The above describes the method for maintaining the cleaning performance of cutting fluid in the embodiments of the present invention. The following describes the device for maintaining the cleaning performance of cutting fluid in the embodiments of the present invention. Please refer to [link / reference]. Figure 2 One embodiment of the cutting fluid cleaning performance maintenance device of the present invention includes: Parameter monitoring module 201: Used to acquire the initial multi-source physical dataset of the cutting fluid; Data preprocessing module 202: used to preprocess the initial multi-source physical dataset to obtain the target multi-source physical dataset; Safety performance judgment module 203: used to call a pre-trained cutting fluid performance judgment model to perform safety performance judgment on the target multi-source physical dataset and obtain judgment results, the judgment results including safety judgment results and warning levels; Instruction generation module 204: used to generate an execution instruction based on the warning level when the security determination result is a non-security level; Instruction execution module 205: used to inject additives into the cutting fluid based on the execution instructions in order to maintain the cleaning performance of the cutting fluid.

[0049] Based on the same ideas as the methods in the above embodiments, the apparatus provided in this application can implement the methods in the above embodiments.

[0050] above Figure 2 The cutting fluid cleaning performance maintenance device in this embodiment of the invention is described in detail from the perspective of modular functional entities. The cutting fluid cleaning performance maintenance equipment in this embodiment of the invention is described in detail below from the perspective of hardware processing.

[0051] Figure 3 This is a schematic diagram of a cutting fluid cleaning performance maintenance device 300 provided in an embodiment of the present invention. The cutting fluid cleaning performance maintenance device 300 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 310 (e.g., one or more processors) and a memory 320, and one or more storage media 330 (e.g., one or more mass storage devices) storing application programs 333 or data 332. The memory 320 and storage media 330 can be temporary or persistent storage. The program stored in the storage media 330 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the cutting fluid cleaning performance maintenance device 300. Furthermore, the processor 310 may be configured to communicate with the storage media 330 and execute the series of instruction operations in the storage media 330 on the cutting fluid cleaning performance maintenance device 300 to implement the steps of the cutting fluid cleaning performance maintenance method provided in the above-described method embodiments.

[0052] The cutting fluid cleaning performance maintenance device 300 may also include one or more power supplies 340, one or more wired or wireless network interfaces 350, one or more input / output interfaces 360, and / or one or more operating systems 331, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 3 The structure of the cutting fluid cleaning performance maintenance device shown does not constitute a limitation on the cutting fluid cleaning performance maintenance device. It may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0053] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the cutting fluid cleaning performance maintenance method.

[0054] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the system, device, or unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0055] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0056] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for maintaining the cleaning performance of cutting fluid, characterized in that, include: Obtain the initial multi-source physical dataset of the cutting fluid; The initial multi-source physical dataset is preprocessed to obtain the target multi-source physical dataset; The pre-trained cutting fluid performance judgment model is invoked to perform a safety performance judgment on the target multi-source physical dataset, and the judgment result is obtained, which includes a safety judgment result and a warning level. When the security determination result is a non-security level, an execution instruction is generated based on the warning level; Additives are injected into the cutting fluid based on the execution command to maintain the cleaning performance of the cutting fluid.

2. The method for maintaining the cleaning performance of cutting fluid according to claim 1, characterized in that, The initial multi-source physical dataset includes a subset of interface performance parameters; The initial multi-source physical dataset for obtaining the cutting fluid includes: The surface tension of the cutting fluid was measured using an industrial online sensor to obtain surface tension parameters. A camera is used to capture images of the morphology of the cutting fluid droplets in the observation window; The contact angle parameters are obtained by performing contact angle calculations based on the morphological image using an image analysis tool. By integrating the surface tension parameter and the contact angle parameter, a subset of interface performance parameters is obtained.

3. The method for maintaining the cleaning performance of cutting fluid according to claim 2, characterized in that, The initial multi-source physical dataset also includes a subset of auxiliary parameters; after integrating the surface tension parameter and the contact angle parameter to obtain the subset of interface performance parameters, it further includes: The conductivity of the cutting fluid was measured using a conductivity sensor to obtain conductivity parameters; The pH sensor was used to measure the acidity and alkalinity of the cutting fluid to obtain the acidity and alkalinity parameters. The turbidity of the cutting fluid was measured using a turbidity sensor to obtain the turbidity parameters. The cutting fluid was subjected to refractive index measurement using a refractometer to obtain refractive index parameters. The auxiliary parameter subset is obtained by integrating the conductivity parameter, the acidity / alkalinity parameter, the liquid turbidity parameter, and the refractive index reading parameter.

4. The method for maintaining the cleaning performance of cutting fluid according to claim 1, characterized in that, The step of preprocessing the initial multi-source physical dataset to obtain the target multi-source physical dataset includes: The initial multi-source physical dataset is processed to unify the units using a physical quantity conversion algorithm to obtain a unified physical dataset. The unified physical dataset is filtered using a moving average algorithm to obtain a filtered physical dataset. The filtered physical dataset is standardized using a standardization algorithm to obtain the target multi-source physical dataset.

5. The method for maintaining the cleaning performance of cutting fluid according to claim 2, characterized in that, The cutting fluid performance judgment model includes a first threshold comparison module, a second threshold comparison module, a performance safety judgment module, and a warning level evaluation module, which are sequentially connected. The pre-trained cutting fluid performance judgment model is used to perform a safety performance judgment on the target multi-source physical dataset to obtain a judgment result. The judgment result includes a safety judgment result and a warning level, including: Obtain the surface tension threshold stored in the first threshold comparison module, and compare the surface tension threshold with the surface tension parameter based on the first threshold comparison module to obtain a first comparison result; Obtain the contact angle threshold stored in the second threshold comparison module, and compare the contact angle threshold with the contact angle parameter based on the second threshold comparison module to obtain a second comparison result; Based on the performance safety determination module, the cutting fluid is subjected to performance safety determination according to the first comparison result and the second comparison result to obtain the safety determination result; When the safety determination result is a non-safety level, the warning level assessment module evaluates the safety performance of the cutting fluid based on the first comparison result and the second comparison result to obtain the warning level.

6. The method for maintaining the cleaning performance of cutting fluid according to claim 5, characterized in that, When the safety determination result is a non-safety level, the warning level assessment module performs a warning level assessment on the safety performance of the cutting fluid based on the first comparison result and the second comparison result to obtain the warning level, including: When the safety determination result is a non-safety level, the number of parameters exceeding the threshold, the surface tension deviation from the threshold, and the contact angle deviation from the threshold are obtained based on the first comparison result and the second comparison result, respectively. Based on the aforementioned warning level assessment module, the safety performance of the cutting fluid is assessed for a warning level according to the number of parameters exceeding the threshold, the surface tension deviation from the threshold parameter, and the contact angle deviation from the threshold parameter, thereby obtaining the warning level.

7. The method for maintaining the cleaning performance of cutting fluid according to claim 1, characterized in that, When the security determination result is a non-security level, an execution instruction is generated based on the warning level, including: When the safety determination result is a non-safety level, an additive selection algorithm is used to perform additive selection processing according to the warning level to obtain the target additive type; A quantitative adjustment algorithm is used to calculate the additive dosage based on the warning level and the target additive type, so as to obtain the compensation dosage corresponding to the target additive type. The execution instructions are generated based on the target additive type and the compensation dose.

8. A device for maintaining the cleaning performance of cutting fluid, characterized in that, include: Parameter monitoring module: used to acquire the initial multi-source physical dataset of the cutting fluid; Data preprocessing module: used to preprocess the initial multi-source physical dataset to obtain the target multi-source physical dataset; Safety performance judgment module: used to call a pre-trained cutting fluid performance judgment model to perform safety performance judgment on the target multi-source physical dataset and obtain the judgment result, which includes safety judgment result and warning level; Instruction generation module: used to generate an execution instruction based on the warning level when the security determination result is a non-security level; Instruction execution module: used to inject additives into the cutting fluid based on the execution instructions in order to maintain the cleaning performance of the cutting fluid.

9. A cutting fluid cleaning performance maintenance device, characterized in that, The cutting fluid cleaning performance maintenance device includes: a memory and at least one processor, wherein the memory stores instructions; At least one of the processors invokes the instructions in the memory to cause the cutting fluid cleaning performance maintenance device to perform the steps of the cutting fluid cleaning performance maintenance method as described in any one of claims 1-7.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instructions are executed by the processor, they implement each step of the cutting fluid cleaning performance maintenance method as described in any one of claims 1-7.

Images