Cutting fluid multi-parameter self-adaptive monitoring and early warning method and system based on dynamic threshold value

Abstract

The invention relates to the technical field of industrial information and data processing, in particular to a cutting fluid multi-parameter self-adaptive monitoring and early warning method and system based on a dynamic threshold value, and aims to obtain and standardize multi-modal data of a cutting fluid in real time, construct a coupling attenuation prediction model in combination with processing working condition parameters, and predict a joint attenuation trend under a specific working condition. A cutting fluid performance time sequence is modeled through a Gaussian process, a confidence interval boundary is dynamically adjusted, and a self-adaptive single parameter dynamic threshold value is calculated. And a Copula function is used to learn a non-linear dependence structure among multiple parameters, a combined overrun risk is identified, and a multi-parameter combination threshold value is dynamically adjusted according to a preset rule. The system monitors data in real time and triggers early warning when abnormity is found, a digital computer is used for intelligent decision support processing, a maintenance knowledge base and a process optimization rule are combined, and accurate maintenance suggestions are generated. In addition, the edge intelligent unit ensures the independent operation capability of the system when the communication is interrupted.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial information and data processing, and specifically to a multi-parameter adaptive monitoring and early warning method and system for cutting fluid based on dynamic thresholds. Background Art

[0002] The severe challenges faced by industrial monitoring and early warning of cutting fluid lie in the deficiencies of digital data processing technology. Specifically, when dealing with massive multi-source heterogeneous data, existing technologies generally have the problem of low data preprocessing efficiency, making it difficult to clean, denoise, and unify the format of real-time high-concurrency sensor data in a timely and effective manner, which directly affects the accuracy of subsequent analysis.

[0003] In addition, after the data is preliminarily processed, the feature extraction and data fusion capabilities are also severely insufficient. Traditional methods are difficult to automatically identify and extract effective features reflecting the key performance of cutting fluid from complex data streams, resulting in the omission of deep correlation information between data. Fixed thresholds cannot achieve precise monitoring and prediction of the state of cutting fluid, and are significantly limited in deeply mining the complex non-linear relationship between the macroscopic state of cutting fluid and the attenuation of its microscopic performance, making it difficult to construct a digital model that can accurately reflect its internal mechanism. Summary of the Invention

[0004] The present invention provides a multi-parameter adaptive monitoring and early warning method and system for cutting fluid based on dynamic thresholds. By deeply integrating multi-modal data from industrial information and data processing with machining condition parameters, a cutting fluid coupling attenuation prediction model is constructed. A single-parameter adaptive threshold is dynamically generated using Gaussian processes, and the Copula function is used to identify multi-parameter non-linear dependencies and joint over-limit risks, and then the multi-parameter combination threshold is dynamically adjusted. The system monitors the state of the cutting fluid in real time, triggers an early warning once the threshold is exceeded, and uses intelligent decision support based on industrial information and data processing to provide accurate maintenance and process adjustment suggestions. It significantly improves the accuracy, foresight, and robustness of early warning, providing strong support for the intelligent management and production optimization of cutting fluid.

[0005] The multi-parameter adaptive monitoring and early warning method for cutting fluid based on dynamic thresholds includes: Real-time collecting multi-modal data of cutting fluid. For data with different sampling frequencies in the multi-modal data, Gaussian processes are used for data interpolation and time alignment processing to generate standardized multi-modal data; Combining the standardized multi-modal data, historical maintenance data, and current working condition parameters to construct a cutting fluid coupling attenuation prediction model to obtain the future state of the cutting fluid; According to the future state of the cutting fluid, using Gaussian processes to model the time series of cutting fluid performance and dynamically adjust the confidence interval boundary, and calculating the adaptive single-parameter dynamic threshold of the key performance indicators of the cutting fluid; Learn the non - linear dependence structure between multi - modal data using Copula function, identify the multi - parameter joint over - limit risk, combine the interaction rules between parameters, and dynamically adjust the multi - parameter combination threshold according to the preset propagation path; Set the trigger warning conditions according to the real - time monitored multi - modal data, perform intelligent decision - making support processing using a digital computer, provide users with intelligent maintenance suggestions including adjusting, supplementing, and replacing cutting fluid, and adjust the process.

[0006] Preferably, the cutting fluid coupling attenuation prediction model includes: Take the current and predicted machining condition parameters as inputs, deduce the mutual influence relationship between the performance indicators of the cutting fluid and the joint attenuation trend under specific condition combinations, and predict the evolution trajectory of each performance indicator.

[0007] Preferably, the calculation of the single - parameter dynamic threshold includes: Based on the output results of the cutting fluid coupling attenuation prediction model, through industrial information and data processing, use the Gaussian process non - parametric probability model to model the time - series data of the cutting fluid performance; Give the confidence interval according to the Gaussian process prediction mean and variance, and calculate the single - parameter dynamic threshold of the key performance indicators of the cutting fluid by dynamically adjusting the boundaries of the confidence interval.

[0008] Preferably, the multi - parameter combination threshold includes: Based on the historical statistical distribution of the standardized multi - modal data and incorporating the output results of the cutting fluid coupling attenuation prediction model, calculate the single - parameter dynamic threshold; Use the Copula function to learn the non - linear dependence structure between the standardized multi - modal data to identify the risk of multi - parameter joint over - limit, perform non - linear combination calculation on the key performance indicators of the cutting fluid, and dynamically adjust the multi - parameter combination threshold according to the preset propagation path.

[0009] Preferably, the intelligent decision - making support processing includes: Use the transfer learning method to initialize and continuously optimize the cutting fluid coupling attenuation prediction model with the model parameters pre - trained on the corresponding equipment, working conditions, and cutting fluid types; Based on the prediction model of the multi - performance coupling attenuation of the cutting fluid and the prediction and analysis results of the machining quality, generate specific suggestions for adjusting the cutting fluid parameters and trigger the replenishment operation of the cutting fluid additives.

[0010] Preferably, the intelligent decision - making support processing further includes: When the system triggers an early warning, based on the type of deterioration parameter, the degree of deterioration, and the remaining service life predicted by the attenuation prediction model as query conditions, match the maintenance plan in the maintenance knowledge base, and combine historical maintenance records. After comprehensive evaluation, generate specific maintenance operation suggestions including accurate supplementary dosage, adjusted parameter range, and recommended replacement time window; Generate suggestions for adjusting machining process parameters based on a preset process optimization rule base.

[0011] A multi-parameter adaptive monitoring and early warning system for cutting fluid based on dynamic thresholds, including: A cutting fluid parameter acquisition and processing module, which is used to collect multi-modal data of cutting fluid in real time. For data with sampling frequency differences in the multi-modal data, Gaussian process is used for data interpolation and time alignment processing to generate standardized multi-modal data; A cutting fluid attenuation prediction module, which is used to combine standardized multi-modal data, historical maintenance data, and current working condition parameters to construct a cutting fluid attenuation prediction model. According to the future state of the cutting fluid, model the time series of cutting fluid performance through Gaussian process and dynamically adjust the confidence interval boundary, and calculate the adaptive single-parameter dynamic threshold of the key performance index of the cutting fluid; a dynamic threshold discrimination module, which is used to learn the non-linear dependence structure between multi-modal data by using the Copula function, identify the risk of joint over-limit of multi-parameters, and combine the interaction rules between parameters to dynamically adjust the multi-parameter combination threshold according to the preset propagation path; An early warning and decision-making module, which is used to set early warning conditions according to the multi-modal data monitored in real time, perform intelligent decision-making support processing by using a digital computer, and provide users with intelligent maintenance suggestions including adjusting, supplementing, and replacing cutting fluid and adjusting the process. Compared with the prior art, the beneficial effects of the present invention are: 1. This cutting fluid early warning method establishes an adaptive dynamic threshold for a single parameter through Gaussian process, and uses the Copula function to identify the non-linear dependence and joint over-limit risk of multi-parameters. This significantly improves the accuracy and timeliness of early warning, and can effectively avoid false alarms and missed alarms of traditional fixed thresholds. By comprehensively understanding the multi-parameter coupling relationship, this method optimizes the information perception ability of the industrial Internet of Things, extends the service life of cutting fluid in industry, and reduces production risks and costs, providing an intelligent and efficient maintenance plan.

[0012] 2. By deeply integrating machining working condition parameters into the cutting fluid coupling attenuation prediction model, this method can deduce the mutual influence between various performance indicators and the joint attenuation trend under specific working conditions, so as to achieve a more comprehensive and in-depth understanding of the health status of cutting fluid. This ability to predict the evolution of multi-performance coupling adaptively based on working conditions provides strong technical support for more reliable early warning, more refined life assessment, and more intelligent maintenance decision-making in industrial Internet of Things information services.

[0013] 3. Traditional monitoring systems usually only provide basic alarms and lack targeted maintenance suggestions. By maintaining a knowledge base and a process optimization rule base, and combining with the prediction of the performance attenuation of cutting fluid, the present invention generates accurate maintenance strategies, such as optimizing the replenishment plan and adjusting machining parameters. This system can not only extend the service life of cutting fluid and tools, but also reduce machining defects caused by deteriorated states, and maintain industrial data integration services, thereby improving the overall production quality and efficiency. In the field of high-end manufacturing, this ability is particularly important, which can significantly reduce the risk of unplanned downtime and improve the stability of the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 It is a flowchart of a multi-parameter adaptive monitoring and early warning method for cutting fluid based on dynamic thresholds proposed in an embodiment of the present invention application; Figure 2 It is a schematic structural diagram of a multi-parameter adaptive monitoring and early warning system for cutting fluid based on dynamic thresholds proposed in an embodiment of the present invention application; Figure 3 It is an interaction process diagram of a multi-parameter adaptive monitoring and early warning method and system for cutting fluid based on dynamic thresholds proposed in an embodiment of the present invention application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0015] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0016] Embodiment 1: Please refer to Figures 1 to 3 , the present invention provides a multi-parameter adaptive monitoring and early warning method and system for cutting fluid based on dynamic thresholds, and the technical solutions are as follows: The multi-parameter adaptive monitoring and early warning method for cutting fluid based on dynamic thresholds includes: S10. Real-time collect multi-modal data of cutting fluid. For the data with different sampling frequencies in the multi-modal data, use Gaussian process for data interpolation and time alignment processing to generate standardized multi-modal data; S20. Combine the standardized multi-modal data, historical maintenance data and current working condition parameters to construct a cutting fluid attenuation prediction model; S30. According to the output of the cutting fluid coupling attenuation prediction model, use Gaussian process to model the time series of cutting fluid performance and dynamically adjust the confidence interval boundary, and calculate the adaptive single-parameter dynamic threshold of the key performance index of cutting fluid; S40. Use the Copula function to learn the non-linear dependence structure among multiple parameters of the cutting fluid, identify the risk of combined over-limit of multiple parameters, and combine the interaction rules among the parameters to dynamically adjust the multi-parameter combination threshold according to the preset propagation path; S50. Compare the state parameters of the cutting fluid monitored in real time with the multi-parameter combination threshold. When the detection exceeds the threshold, trigger the warning condition, and use a digital computer for intelligent decision support processing to provide users with intelligent maintenance suggestions including adjusting, replenishing, and replacing the cutting fluid and adjusting the process.

[0017] As an implementation manner of the present invention, refer to Figure 1 , the flow chart of the multi-parameter adaptive monitoring and warning method for cutting fluid based on dynamic threshold.

[0018] Furthermore, the multi-modal data acquisition method in the industrial database is as follows: At key nodes of the cutting fluid circulation system, including the supply pipeline, return pipeline, and liquid tank, 3 independent online sensor technologies are integrated to collect multi-modal data. Among them, the physical and chemical indexes of the cutting fluid are monitored by microscopic sensors including pH, conductivity, turbidity, refractometer, absorbance, and temperature; the biosensor detects the volatile products generated by microbial metabolism in real time to evaluate biological contamination; in addition, the electrochemical ORP sensor indirectly reflects the microbial activity through the change of redox potential, and combines the principle of impedance microbiology to comprehensively and dynamically monitor the health status of the cutting fluid. After the multi-modal data acquisition is completed, first uniformly convert the physical units of the data of each sensor, and map the multi-source data to a unified dimension through normalization; given the state parameters of the cutting fluid at the current moment as the initial condition, use the numerical integration method and the current and predicted working condition parameters as input items to predict the change trend of these key indexes within the next 8 to 24 hours.

[0019] Furthermore, the specific steps for the deployment and execution of the online parameter estimation algorithm are to select the EKF parameter estimation algorithm, recursively update, and based on the model parameters and states estimated at the previous moment, use the hybrid model to predict the state variables at the current moment to determine the unified alignment time point t_align.

[0020] Periodically generate the target alignment time point, and perform alignment and interpolation for each t_align: Synchronize high-frequency data (pH, temperature, RPM): Directly extract the latest data points at or closest to the t_align moment from their respective buffers.

[0021] Asynchronous data (turbidity, ORP): Extract all historical timestamp and corresponding value data points near t_align from the sliding window buffer corresponding to the asynchronous sensor. Using the extracted timestamps as the input and the corresponding values as the output, dynamically construct a Gaussian process regression model for this sensor to model the smooth trend and noise. Use the trained Gaussian regression model to make a prediction at the target time point t_align to obtain the interpolation value of this sensor.

[0022] Perform standardized processing of unified dimension for all aligned numerical data to generate standardized multimodal data.

[0023] Furthermore, directly read the current spindle speed, table and tool post movement speed, and cutting depth through the machine tool numerical control system. Obtain the production plan for the next period of time from the APS advanced planning and scheduling system, and parse out the workpiece material, tool type, and parameters in the preset machining program corresponding to the planned machining tasks. In addition, the workpiece material and tool type are used as attribute inputs associated with specific machining tasks. Numerically encode the workpiece material and tool type parameters, and standardize the continuous parameters of speed, feed, and cutting depth so that they are comparable with the cutting fluid's own parameters in the numerical range.

[0024] Take the processed working condition parameters, together with the cutting fluid's own multimodal parameters and historical data, as the input features of the Transformer coupled attenuation prediction model.

[0025] Use the key performance indicators, key chemical component concentrations, and working condition parameters of the cutting fluid as nodes. The edges represent the direct influence relationships between the indicators. There is an edge between the lubricity node and the cooling node, between the micro-index node and the pH node and the rust prevention node, and between the working condition parameter node and all performance nodes. The weights of the edges are obtained through presetting.

[0026] Furthermore, the input sequence of the sequence modeling process is set to include the multi-dimensional performance state of the cutting fluid at historical moments, the corresponding working condition parameters, and timestamp information. The attention mechanism is set so that the model learns which moments' states and working condition combinations in the historical sequence should be focused on when predicting future performance, as well as the mutual influence between different performance indicators.

[0027] Use the historical data including the cutting fluid multi-parameter monitoring data, the corresponding working condition data, and the known performance attenuation results to conduct supervised learning training on the Transformer model. The loss function is set as the weighted sum of the prediction errors of multiple performance indicators. Use the trained model to make a joint attenuation trend prediction. After receiving the current cutting fluid state and the current and future predicted working condition parameters, it can simultaneously output the evolution trajectory of the cutting fluid's key performance indicators within the next 8 to 24 hours, that is, the predicted value sequence.

[0028] By deeply integrating the machining condition parameters into the coupling attenuation model and using advanced algorithms to deduce the depth of the condition parameters integrated into the cutting fluid coupling attenuation prediction model, the mutual influence between various performance indicators and the joint attenuation trend under specific working conditions are deduced, achieving a more comprehensive and in-depth understanding of the health status of the cutting fluid. It can also provide strong technical support for more reliable early warning, more refined life assessment, and more intelligent maintenance decisions based on the leap of condition-adaptive multi-performance coupling evolution prediction.

[0029] Furthermore, historical time series data of specific parameters in the healthy state are collected. Calculate the mean and standard deviation of each parameter and fit its probability distribution to set a benchmark threshold range: , where, is the parameter mean, is the standard deviation, and are coefficients selected according to the risk tolerance and parameter characteristics.

[0030] The dynamic threshold of a single parameter is the result of the combined action of the benchmark threshold, historical statistical distribution characteristics, and future short-term prediction trends. Let the predicted value obtained at the th sampling period at the current time be , and the dynamic lower limit can be expressed as: ; where, is the weight factor, represents the maximum value, is the current value of this parameter, is the set absolute safety lower limit. In this calculation method, when the prediction rate drops rapidly, the dynamic threshold will be triggered earlier than the static threshold.

[0031] Collect historical data points of specific performance indicators from the cutting fluid sensor or laboratory analysis data in chronological order. According to the characteristics of the cutting fluid performance time series, select the radial basis function to capture the main trend and supplement it with the white noise kernel to simulate the measurement error. Subsequently, through the maximum likelihood estimation optimization method, continuously learn and adjust the hyperparameters of the kernel function to ensure that the model can fit the historical data to the greatest extent.

[0032] Furthermore, the Gaussian process model is trained using the preprocessed cutting fluid performance time series data. The training process includes calculating the covariance matrix and learning the hyperparameters of the model. The trained Gaussian process model is used to predict the cutting fluid performance indicators in the short term future. The Gaussian process can give the prediction mean and prediction variance, based on which the confidence interval is calculated. For the performance indicators fluctuating within a certain range, the dynamic threshold can simultaneously include the upper and lower limits of the adjusted confidence interval to form the normal interval. Set the initial confidence level as , let be the mean, be the variance, and the confidence interval is: ; where is the Z value corresponding to the confidence level in the standard normal distribution.

[0033] Based on the prediction results of the cutting fluid performance decay, the system dynamically adjusts the asymmetric confidence interval: when an accelerating decay trend is detected, the warning lower limit is adjusted closer to the prediction mean to achieve early warning; at the same time, the boundary width is automatically adjusted according to the historical data fluctuation range, tightened when stable and widened when fluctuating greatly, and continuously optimized through false alarm feedback to form an adaptive dynamic threshold warning mechanism.

[0034] This cutting fluid monitoring and warning method models the performance decay using the Gaussian process non-parametric probability model, dynamically adjusts the single-parameter confidence interval by combining the short-term future trend prediction to generate an adaptive threshold, and at the same time deeply learns and identifies the non-linear dependence structure among multiple parameters of the cutting fluid with the help of the Copula function, realizing the accurate identification of the multi-parameter joint overlimit risk and the dynamic threshold adjustment. At the same time, through the comprehensive insight into the multi-parameter coupling relationship, the overall health status of the cutting fluid can be evaluated more comprehensively, the cutting fluid management can be optimized, the service life can be extended, and a more intelligent and efficient cutting fluid maintenance solution can be provided.

[0035] The Copula function can connect multiple one-dimensional marginal distribution functions to construct its joint distribution function, that is, any dimensional joint distribution function can be expressed by a dimensional Copula function and its marginal distribution functions as , where is the standardized multimodal data; By constructing the joint distribution, the dependence structure of multiple variables is separated from their respective marginal distributions for modeling. By taking the partial derivative of the joint cumulative distribution function, the joint probability that the standardized multimodal data is in an adverse state simultaneously is obtained. The multi-parameter combination threshold is determined, the overrun area is defined, and the risk of multi-parameter joint overrun is identified. In this embodiment, the joint probability of entering this predefined "overrun area" within the next 3 sampling periods is 0.3647, while the set risk threshold is 0.10. Because , a multi-parameter combination warning is triggered.

[0036] Table 1: Construction of Rule-Based Decision Logic

[0037] To more precisely capture the non-linear dependence structure that dynamically changes over time among the multi-parameter data of the cutting fluid and improve the robustness of multi-parameter joint overrun risk identification, this method deeply optimizes the selection of the Copula function, parameter learning, and data processing. The performance degradation of the cutting fluid is a dynamic process, and the dependence structure among its parameters may also evolve over time. In the initial wear stage and the later exhaustion stage, the mutual influence among parameters may be different. To capture this dynamicity, the following strategy is adopted for parameter learning of the Copula function: at each preset update interval, only the data within the current rolling window is used, and the Copula parameters are gradually updated based on stochastic gradient descent to obtain the self-updating Copula family parameters and their maximum likelihood estimates, ensuring that the Copula model always reflects the latest changes in the dependence structure and timely adjusting the assessment of multi-parameter joint risk.

[0038] By incorporating future short-term trend prediction to calculate the multi-parameter combination threshold, the warning is no longer based solely on the current values, but can "foresee" the overrun risk that the cutting fluid will encounter. Through the predicted decay of the lubricity index LI and the increase in surface roughness Ra, the system can issue a warning in advance, which buys valuable reaction time for the operator. It implies that the decrease in lubricity leads to an increase in surface roughness, thus providing deeper insights for fault diagnosis and formulating refined predictive maintenance strategies, providing deeper insights for fault diagnosis and maintenance decision-making, precisely evaluating the joint risk and understanding the deterioration propagation path, formulating more refined predictive maintenance strategies, and effectively avoiding equipment downtime and production losses.

[0039] Furthermore, it is necessary to collect historical monitoring data and performance degradation data from multiple similar devices connected during industrial network operation. This data includes machine tools of different models but sharing the same processing principles, as well as data generated under similar operating conditions (processing the same material with slightly different parameters, or using cutting fluids from the same series but different batches / brands). This historical monitoring and performance degradation data comprehensively characterizes the dynamic changes of cutting fluids during actual industrial network operation, providing critical support for building and validating cutting fluid degradation models.

[0040] This data is aggregated to construct a relatively large dataset of universal cutting fluid performance degradation. The Transformer model is applied to this universal dataset to learn the general patterns of cutting fluid performance degradation and the universal dependencies between parameters. While protecting data privacy, the fine-tuned model parameters across multiple edge devices are aggregated to update the global pre-trained model. This updated model is then distributed to enable continuous collaborative model evolution.

[0041] Transfer learning enables the model to start learning from a better initial state on the target task, especially when the target task data is scarce and difficult to obtain. The universal knowledge learned by the pre-trained model from diverse source data helps improve the model's performance when faced with unseen and slightly different data in the target task, enhancing the model's generalization ability and robustness as well as its adaptability to unexpected situations.

[0042] Furthermore, the maintenance knowledge base construction process includes storing standardized maintenance plans for different cutting fluid degradation parameters and degrees of degradation, including mild, moderate, and severe, as well as varying remaining useful life (RUL) intervals. Each plan details recommended operating procedures, required materials, and precautions. Knowledge is then represented using conditional judgment rules, frameworks, and semantic networks.

[0043] The degradation parameter type and degree are directly obtained from the early warning information, and the RUL prediction is obtained from the attenuation prediction model.

[0044] An algorithm that uses keyword matching, fuzzy matching, and vector space models to calculate the similarity between query conditions and knowledge base cases is used to match the most suitable maintenance plan, ensuring that the pushed plan is highly relevant. Comprehensive evaluation and adjustment are carried out in combination with historical maintenance records from the Industrial Internet of Things technology system service. The service stores past maintenance operations through its integrated historical record database, including time, type, specific measures, operators, and feedback on maintenance effects.

[0045] The failure of the cutting fluid is not caused by the extreme abnormality of a single parameter, but by the synergistic effect of the "sub-healthy" states of multiple parameters. Regarding each parameter as a node and the pairwise dependence relationships and interaction rules of the third order and above among the parameters as edges or hyper-edges, a parameter interaction network is constructed, which serves as the core basis for subsequent identification of combined over-limit risks and dynamic adjustment of combined thresholds. When abnormal fluctuations occur in the initial parameters, the network can dynamically track their influence paths - the abnormal signals propagate in the network through edges and hyper-edges, ultimately leading to the chain decline of the key performance indicators of the cutting fluid. When a synergistic effect is formed in a specific manner, the system can timely warn of potential failure risks, providing a more refined decision-making basis for preventive maintenance.

[0046] By solidifying the maintenance experience and best practices of industry experts therein, the system can automatically match and recommend the optimal maintenance plan according to the real-time working conditions, thus realizing the intelligence and standardization of maintenance decision-making, greatly reducing the dependence on individual experience, and ensuring the scientificity and consistency of decisions. On this basis, the system conducts in-depth learning and dynamic adjustment in combination with historical maintenance records, making each maintenance suggestion, whether it is the supplementary dosage or the replacement cycle, more in line with the actual needs of specific equipment and specific working conditions, realizing the precision and personalization of maintenance operations, effectively avoiding the extensive management of "one-size-fits-all", and significantly improving the pertinence and actual effect of maintenance. When a warning event occurs, the system can respond quickly, provide specific and operable maintenance guidance, greatly improving the emergency response ability and problem-solving efficiency, and significantly shortening the fault diagnosis and handling time.

[0047] Table 2: Specific maintenance operation suggestions combined with the maintenance knowledge base

[0048] To achieve reliable monitoring and early warning in the edge computing environment, the system adopts a multi-level data validity verification mechanism, including range check, null value, outlier detection, and change rate check, and marks, eliminates, and interpolates and repairs invalid data. At the feature extraction level, the system calculates time-domain features and frequency-domain features in real time, including spectral energy, main frequency, and derivative features for specific parameters. The lightweight early warning model is deployed with a three-level architecture: simple threshold alarm, rule-based IF-THEN early warning, and decision tree optimized neural network. The edge intelligent unit, as a key component in the industrial Internet of Things architecture, has a complete autonomous operation ability. Through local data storage and processing, and the local deployment of core monitoring logic, it can continue to work offline in a complex industrial Internet of Things environment. When the industrial Internet of Things communication link is interrupted, the edge intelligent unit can automatically switch to the offline mode, trigger early warning by relying on the locally deployed lightweight model and rules through indicator lights and buzzers, and cache all generated monitoring data at the same time. After the industrial Internet network communication is restored, it will synchronize the data to the cloud to ensure data integrity and subsequent advanced analysis.

[0049] This design enables the system to maintain core monitoring and basic early warning functions even when communication is restricted and some components fail, ensuring the integrity of key data and business continuity. The preliminary processing and early warning judgment at the edge end improve the real-time response speed and efficiency of the system, which is particularly crucial for abnormal events that require rapid response and avoid cloud processing delays. At the same time, due to the huge redundancy of raw data and high transmission costs, which cause heavy pressure on cloud processing, edge filtering and feature extraction are adopted to intelligently screen and refine information at the source of data generation, not only improving the response speed and decision-making efficiency of the system, but also maximizing the data value at a lower cost. The autonomous ability further enhances the robustness and usability of the system, ensuring that the core monitoring and basic early warning can still be independently executed when the network is interrupted, and guaranteeing basic safety monitoring. In addition, the source data verification improves the data quality, provides a more reliable input for complex model analysis in the cloud, and thus improves the analysis accuracy. Finally, the rapid response at the edge and the in-depth analysis in the cloud form an efficient complementarity, jointly building a hierarchical intelligent monitoring and early warning system.

[0050] Furthermore, by integrating monitoring methods, a multi-parameter adaptive monitoring and early warning system for cutting fluid based on dynamic thresholds is obtained, including a cutting fluid parameter acquisition and processing module, a cutting fluid attenuation prediction module, a dynamic threshold discrimination module, and an early warning and decision-making module. Refer to Figure 2 This is the structural schematic diagram of the multi-parameter adaptive monitoring and early warning system for cutting fluid based on dynamic thresholds proposed in the embodiment of the present invention application, corresponding to: ​ A cutting fluid attenuation prediction module, which is used to combine standardized multimodal data, historical maintenance data and current working condition parameters to construct a cutting fluid attenuation prediction model, and according to the future state of the cutting fluid, model the time series of cutting fluid performance through Gaussian process and dynamically adjust the confidence interval boundary, and calculate the adaptive single-parameter dynamic threshold of the key performance indicators of the cutting fluid; A dynamic threshold discrimination module, which is used to learn the non-linear dependence structure between multimodal data by using the Copula function, identify the risk of joint over-limit of multiple parameters, and combine the interaction rules between parameters to dynamically adjust the multi-parameter combination threshold according to the preset propagation path; An early warning and decision-making module, which is used to set early warning conditions according to the real-time monitored multimodal data, perform intelligent decision-making support processing by using a digital computer, and provide users with intelligent maintenance suggestions including adjusting, replenishing and replacing cutting fluid and adjust the process.

[0051] This system realizes the comprehensive perception of the cutting fluid state through multimodal data fusion technology; constructs a dynamic threshold adaptive mechanism based on the Transformer prediction model and the Copula function, which can accurately identify the risk of joint deterioration of multiple parameters; adopts an edge-cloud collaborative architecture to ensure real-time response ability through lightweight edge computing, and at the same time relies on cloud big data to optimize long-term prediction performance; finally forms a complete intelligent decision-making closed loop from state monitoring, intelligent early warning to maintenance suggestion generation, and realizes automated maintenance decision-making driven by the whole process data.

[0052] The present invention establishes an adaptive dynamic threshold for a single parameter through Gaussian process, and identifies the non-linear dependence and joint over-limit risk of multiple parameters by means of the Copula function, which significantly improves the accuracy and timeliness of cutting fluid early warning and effectively avoids false alarms and missed alarms of traditional fixed thresholds. Deeply integrating the processing working condition parameters, it realizes the comprehensive insight into the health status of the cutting fluid and the coupling attenuation prediction under specific working conditions. Cooperating with the intelligent decision-making support system to provide precise maintenance and process adjustment suggestions, and the edge intelligent unit to ensure the robustness of the system, the present invention provides a powerful and intelligent technical support for the lean management, life extension and production optimization of cutting fluid, and greatly reduces the operation cost and production risk.

[0053] At the same time, this method provides highly customized and highly operable intelligent maintenance decision-making support, provides specific and effective data for users, thereby extending the effective life of the cutting fluid, reducing the working condition adaptability and multi-performance coupling analysis ability of the professional system of the operator and the maintenance cost, and improving the production efficiency.

[0054] Example Two: This embodiment uses high-precision titanium alloy (such as Ti-6Al-4V) milling. Titanium alloy is a difficult-to-machine material, which has extremely high requirements for the cooling, lubrication, and chemical stability of cutting fluid, and it is easy to chemically react with cutting fluid. It is commonly used in aerospace parts with very high machining accuracy requirements.

[0055] It is necessary to simultaneously pay attention to lubricity, cooling efficiency, and chemical stability, including specific corrosion inhibitor concentration, titanium ion dissolution concentration, and antioxidant properties. The coupling effects among these properties are more prominent in the machining of titanium alloys with high temperature and high chemical activity.

[0056] Introduce the dynamic assessment of the remaining useful life (RUL) of the cutting fluid under the current and future production schedules. The maintenance decision is not just a single top-up, but a phased maintenance strategy based on RUL and performance prediction, including minor adjustments, enhanced top-ups, planned replacements, and may trigger coordinated adjustments of process parameters to complete urgent tasks without immediately replacing the cutting fluid.

[0057] The maintenance knowledge base and rules are more refined and require more factors to be considered. The knowledge base needs to contain knowledge about specific additives for titanium alloy machining, including high-temperature lubricants, titanium corrosion inhibitors, and antioxidants. Their dose-response relationships may be more complex and affected by other parameters. The decision-making needs to consider maintenance costs, downtime, and the consumption of cutting fluid performance by future production plans.

[0058] Therefore, in the machining of titanium alloys, the multi-performance coupling of cutting fluid needs to be monitored synergistically, and the RUL is dynamically predicted in combination with the attenuation characteristics under high-temperature and high-activity working conditions. Based on RUL and performance coupling effects, a phased maintenance strategy is formulated, and the dose-response knowledge base of special additives for titanium alloys is integrated. The maintenance costs, downtime, and production plans are comprehensively weighed to achieve dynamic optimization decisions.

[0059] Furthermore, according to RUL and performance prediction, a phased maintenance strategy is adopted, including minor adjustments, enhanced top-ups, planned replacements, and trigger coordinated adjustments of process parameters to complete urgent tasks without immediately replacing the cutting fluid. The entire maintenance process relies on a continuously updated cutting fluid performance database, and the judgment thresholds and response strategies for each stage are continuously optimized through the Transformer model to continuously improve the maintenance efficiency.

[0060] By systematically integrating the key parameter indicators of the cutting fluid, their corresponding threshold standards, the current values of real-time monitoring, and the future trend prediction based on the model, the targeted maintenance measures and optional process adjustment suggestions generated by the intelligent system are finally clearly presented, thus transforming the complex monitoring and decision-making process into an intuitive and easy-to-understand action guide to achieve intelligent management of the cutting fluid.

[0061] Table 3: Intelligent Monitoring and Maintenance Decision

[0062] "Intelligent Monitoring and Maintenance Decision-Making" intuitively presents the system's real-time monitoring results of key cutting fluid parameters and the corresponding maintenance strategies. The core technical foundation supporting these intelligent decisions is reflected in the predictive model and rule base. This forms a complete closed loop of intelligent management of industrial information and data processing: advanced algorithmic models provide a scientific basis for decision-making, and execution feedback continuously optimizes the model accuracy in the rule base. This synergistic mechanism of data-driven, model-supported, and decision-feedback is the core advantage of this system in achieving predictive maintenance.

[0063] Reference Figure 3 , which is an interactive process diagram of the cutting fluid multi-parameter adaptive monitoring and early warning method and system based on dynamic thresholds proposed in the embodiment of the application of the present invention.

[0064] Table 4: Prediction model and rule base summary

[0065] This example clearly illustrates how the present invention provides industrial data integration services under demanding and complex working conditions. It demonstrates how, through multi-parameter, multi-model, and multi-stage intelligent decision-making, it achieves refined management of cutting fluids, predictive maintenance, and effective collaboration with production processes. This improves working condition adaptability, deepens performance insights, optimizes maintenance strategies, proactively mitigates risks, and achieves collaborative and economical production and maintenance. This fully demonstrates the significant advantages and advancements of the present invention compared to traditional methods and industrial Internet scenarios.

[0066] 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 multi-parameter adaptive monitoring and early warning method for cutting fluid based on dynamic thresholds, characterized in that, It includes the following steps: Collect multi-modal data of the cutting fluid in real time. For the data with different sampling frequencies in the multi-modal data, use Gaussian process for data interpolation and time alignment processing to generate standardized multi-modal data; Combine the standardized multi-modal data, historical maintenance data and current working condition parameters to construct a cutting fluid coupling attenuation prediction model to obtain the future state of the cutting fluid; According to the future state of the cutting fluid, use Gaussian process to model the time series of the cutting fluid performance and dynamically adjust the confidence interval boundary, and calculate the adaptive single-parameter dynamic threshold of the key performance indicators of the cutting fluid; Use the Copula function to learn the non-linear dependence structure between multi-modal data, identify the risk of multi-parameter joint overrun, and combine the interaction rules between parameters to dynamically adjust the multi-parameter combination threshold according to the preset propagation path; According to the multi-modal data monitored in real time, set the trigger warning conditions, and use a digital computer for intelligent decision support processing to provide users with intelligent maintenance suggestions including adjusting, replenishing and replacing the cutting fluid and adjusting the process.

2. The multi-parameter adaptive monitoring and early warning method for cutting fluid based on dynamic threshold according to claim 1, characterized in that, The cutting fluid coupling attenuation prediction model includes: Take the current and predicted processing condition parameters as inputs, deduce the mutual influence relationship between the performance indicators of the cutting fluid and the joint attenuation trend under the preset condition combination, and predict the evolution trajectory of each performance indicator.

3. The multi-parameter adaptive monitoring and early warning method for cutting fluid based on dynamic threshold according to claim 1, wherein The calculation of the single-parameter dynamic threshold includes: Based on the output result of the cutting fluid coupling attenuation prediction model, through industrial information and data processing, use the Gaussian process non-parametric probability model to model the time series data of the cutting fluid performance to obtain the Gaussian process prediction mean and variance; Give the confidence interval according to the Gaussian process prediction mean and variance, and calculate the single-parameter dynamic threshold of the key performance indicators of the cutting fluid by dynamically adjusting the boundary of the confidence interval.

4. The multi-parameter adaptive monitoring and early warning method for cutting fluid based on dynamic threshold according to claim 1, characterized in that The multi-parameter combination threshold includes: Based on the historical statistical distribution of the standardized multi-modal data and incorporate the output result of the cutting fluid coupling attenuation prediction model; Use the Copula function to learn the non-linear dependence structure between the standardized multi-modal data to identify the risk of multi-parameter joint overrun, perform non-linear combination calculation on the key performance indicators of the cutting fluid, and dynamically adjust the multi-parameter combination threshold according to the preset propagation path.

5. The multi-parameter adaptive monitoring and early warning method for cutting fluid based on dynamic threshold according to claim 1, wherein, The intelligent decision support processing includes: Use the transfer learning method to use the model parameters pre-trained on the corresponding equipment, working conditions and cutting fluid types for the initialization and continuous optimization of the cutting fluid coupling attenuation prediction model; Based on the prediction model of the multi-performance coupling attenuation of the cutting fluid and the processing quality prediction and analysis results, generate specific suggestions for adjusting the multi-modal data of the cutting fluid and trigger the replenishment operation of the cutting fluid additive.

6. The multi-parameter adaptive monitoring and early warning method for cutting fluid based on dynamic threshold according to claim 5, characterized in that, The intelligent decision support processing also includes: When the system triggers a warning, use the type of deterioration parameter, the degree of deterioration and the remaining service life predicted by the attenuation prediction model as query conditions to match the maintenance plan in the maintenance knowledge base, and combine the historical maintenance records to comprehensively evaluate and generate specific maintenance operation suggestions including the accurate replenishment dose, the adjustment parameter range and the recommended replacement time window. Generate suggestions for adjusting machining process parameters based on a preset process optimization rule base.

7. The multi-parameter adaptive monitoring and early warning system for cutting fluid based on dynamic threshold is characterized in that Including: A cutting fluid parameter acquisition and processing module, which is used to collect multi-modal data of the cutting fluid in real time. For data with different sampling frequencies in the multi-modal data, Gaussian process is used for data interpolation and time alignment processing to generate standardized multi-modal data; A cutting fluid attenuation prediction module, which is used to combine the standardized multi-modal data, historical maintenance data and current working condition parameters to construct a cutting fluid attenuation prediction model. According to the future state of the cutting fluid, Gaussian process is used to model the time series of the cutting fluid performance and dynamically adjust the confidence interval boundary, and calculate the adaptive single-parameter dynamic threshold of the key performance index of the cutting fluid; A dynamic threshold discrimination module, which is used to learn the non-linear dependence structure between multi-modal data by using the Copula function, identify the risk of joint over-limit of multiple parameters, and combine the interaction rules between parameters to dynamically adjust the multi-parameter combination threshold according to the preset propagation path; An early warning and decision-making module, which is used to set early warning conditions according to the real-time monitored multi-modal data, and perform intelligent decision support processing by using a digital computer, and provide users with intelligent maintenance suggestions including adjusting, supplementing and replacing the cutting fluid and adjusting the process.

Images