Lubricating oil viscosity monitoring method and system based on parameter coupling and digital twinning

Abstract

The lubricating oil viscosity monitoring method and system based on parameter coupling and digital twinning belong to the technical field of lubricating oil performance monitoring, and comprises the following steps: step S1, collecting viscosity data of lubricating oil under different temperature-shear rate by using a rheometer; step S2, constructing a multi-parameter dynamic coupling viscosity-temperature model; step S3, building a digital twinning platform; step S4, based on the model output curve; step S5, dynamic correction and performance prediction. The temperature-shear rate coupling model is established, the viscosity-temperature curve fitting error is reduced, and the monitoring precision is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of lubricating oil performance monitoring, and particularly relates to a lubricating oil viscosity monitoring method and system based on parameter coupling and digital twinning. BACKGROUND

[0002] In the modern industrial field, the viscosity of lubricating oil as a core index for measuring its lubricating performance is directly related to the friction and wear characteristics and the service life of the mechanical system, and is crucial to the stable and efficient operation of the equipment. From the development of the prior art, the industry has long relied on traditional viscosity-temperature models to fit and predict the viscosity-temperature characteristics of lubricating oil. These models estimate the viscosity of lubricating oil by establishing a correlation between viscosity and temperature, providing a basic method for lubricating performance evaluation.

[0003] However, the traditional viscosity-temperature model generally has insufficient accuracy and poor dynamics, and the core defect is that it only considers the effect of temperature on viscosity and completely ignores the role of shear rate, a key parameter in actual working conditions. Studies have shown that when the shear rate exceeds 6000 s -1 , the lubricating oil will exhibit significant non-Newtonian characteristics, and the viscosity will change dynamically with the shear rate. Since the traditional model does not take this effect into account, the fitting results deviate greatly from the actual working conditions, and cannot reflect the true dynamic change law of viscosity. SUMMARY

[0004] In view of the above shortcomings and deficiencies of the prior art, the application provides a lubricating oil viscosity monitoring method and system based on parameter coupling and digital twinning, which reduces the viscosity-temperature curve fitting error by establishing a temperature-shear rate coupling model, and improves the monitoring accuracy.

[0005] To achieve the above purpose, the main technical solutions adopted by the application include:

[0006] The lubricating oil viscosity monitoring method based on parameter coupling and digital twinning includes the following steps:

[0007] Step S1, a rheometer is used to collect viscosity data of lubricating oil at different temperatures and shear rates, abnormal values are removed, and effective data are retained as model input data for step S2;

[0008] Step S2, a multi-parameter dynamic coupling viscosity-temperature model is constructed as follows:

[0009] μ(T, γ) = a(T)(log γ)³ + b(T)(log γ)² + c(T)log γ + d(T);

[0010] Wherein, μ is the dynamic viscosity, ; T is the temperature, ℃; γ is the shear rate, s -1; a(T), b(T), c(T) are coefficients of log(γ) 3 ; a(T), b(T), c(T) are coefficients of log(γ) 2 ; d(T) is constant term

[0011] a(T), b(T), c(T), d(T) are determined as follows:

[0012] Set η 预测 = a(log(γ))³ + b(log(γ))² + c(log(γ)) + d

[0013] where η 预测 is predicted viscosity, γ is shear rate, a, b, c, d are four parameters

[0014] An adaptive damping factor λ is added to the Levenberg-Marquardt algorithm, and the optimized L-M algorithm is as follows:

[0015] ;

[0016] where J k is the partial derivative matrix of residual to parameters at the kth iteration, that is, the Jacobian matrix; λ k is the adaptive damping factor at the kth iteration, r k is the residual vector at the kth iteration, I is the unit matrix, k is the iteration number, d k is the increment of the parameter vector to be solved at the kth iteration;

[0017] The optimization objective is to minimize the sum of squares of residuals: S(x) =∑[η 实测 - η 预测 ]², where η 实测 is the actually measured viscosity

[0018] The residual ratio is calculated by the following formula:

[0019] ;

[0020] where: is the residual ratio at the kth iteration, is the minimized sum of squares of residuals at the kth iteration, is the minimized sum of squares of residuals at the kth iteration after d k correction;

[0021] λ k is adjusted according to ρ k ​, update the parameter vector x until the residual sum of squares is less than a preset convergence threshold or a maximum number of iterations is reached, obtain the optimal parameters [a, b, c, d] at the current temperature T, and calculate the optimal parameters corresponding to different temperature points , the obtained , , , As a new data set, the parameters a, b, c, and d are determined as functions of temperature T through polynomial fitting a(T), b(T), c(T), and d(T);

[0022] Step S3, build a digital twin platform:

[0023] The digital twin platform is composed of a physical layer, a virtual layer, and a data interaction layer. The physical layer is a temperature sensor, a torque sensor, and a data acquisition card set for a rheometer, which collects the temperature and shear rate-viscosity data of the lubricating oil in real time. The virtual layer is a virtual simulation model built based on the multi-parameter dynamic coupling viscosity-temperature model constructed in step S2, and integrates a sub-model of oil molecule motion under shear action. The data interaction layer realizes real-time data synchronization between the physical layer and the virtual layer using the OPC UA protocol.

[0024] Step S4, based on the model output curve:

[0025] The type of the output curve is determined to be a "viscosity-temperature curve" under a fixed shear rate. Based on the multi-parameter dynamic coupling viscosity-temperature model of step S2, the corresponding dynamic viscosity μ is calculated at the selected shear rate, and the curve of viscosity changing with temperature at this shear rate is obtained. The digital twin platform of step S3 is used to display the curve.

[0026] Step S5, dynamic correction and performance prediction:

[0027] The digital twin platform compares the measured viscosity μ 实测 of the physical layer with the predicted viscosity μ 预测 of the virtual layer in real time, and calculates the relative error:

[0028] ;

[0029] Set an error threshold. When the error exceeds the set threshold, trigger the feedback mechanism, call the optimized L-M algorithm to refit the coefficients a, b, c, and d to correct the model, until the error is not greater than the set threshold. Based on the corrected model, output the viscosity-temperature curves under different temperature and shear rate combinations.

[0030] ​Generate multi-dimensional early warning information: when the viscosity change rate is greater than 10% / min, trigger the "viscosity abnormal fluctuation early warning"; when the temperature T exceeds the effective lubricating temperature range of the lubricating oil, issue a "critical temperature threshold early warning".

[0031] Further, in the step S1, the temperature regulation range of the rheometer covers-40℃~+150℃; during the collection process, the temperature is changed in a gradient at a rate of 5℃ / min in the range of 20℃~80℃, each temperature point is stable for more than 10 minutes, and the shear rate is set to 10~38000s -1 For each temperature-shear rate combination, continuously collect 30-second viscosity data with a sampling interval of 1 second, and take the average value as the effective data.

[0032] Further, in the step S2, according to ρ k Adjust λ k , if 0.75<ρ k <1, adjust λ k+1 =λ k / 10; if 0<ρ k <0.25, adjust λ k+1 =10λ k ; if 0.25≤ρ k ≤0.75, then λ k+1 =λ k ; wherein ρ k is the residual ratio at the kth iteration, λ k is the adaptive damping factor at the kth iteration, and λ k+1 is the adaptive damping factor at the kth+1 iteration.

[0033] Further, in the step S2, if ρ k is greater than 0, update the parameter x k+1 =x k +d k , wherein ρ k is the residual ratio at the kth iteration, x k+1 is the parameter vector at the kth+1 iteration, x k is the parameter vector at the kth iteration, and d k is the increment of the parameter vector to be solved at the kth iteration.

[0034] Further, in the step S3, the data collected by the physical layer of the digital twin platform is transmitted to the edge computing gateway through the industrial bus for filtering, normalization preprocessing and uploading.

[0035] Further, in the step S3, the transmission period of the data interaction layer is ≤100ms.

[0036] Further, in step S3, a real-time database is built under the digital twin platform to store historical data, supporting data backtracking and model iteration.

[0037] Further, in step S4, during the output curve process, the temperature T parameter range is selected as-40℃~+150℃, and the shear rate γ parameter range is selected as 10~38000s -1 .

[0038] Further, when the error exceeds 5%, a feedback mechanism is triggered to call the optimized L-M algorithm to re-fit the coefficients a, b, c, d, until the error is ≤5%, and the model is dynamically updated.

[0039] A lubricating oil viscosity monitoring system based on parameter coupling and digital twinning, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of the lubricating oil viscosity monitoring method based on parameter coupling and digital twinning when executing the computer program.

[0040] The lubricating oil viscosity monitoring method and system based on parameter coupling and digital twinning have the following beneficial effects: the method is based on the improved L-M algorithm, a temperature-shear rate coupling model is created, the viscosity-temperature curve fitting error is less than 5%, and the viscosity data calculation accuracy is improved.

[0041] The three-layer architecture of the digital twin platform is relied on to realize 50ms-level real-time correction and improve the response speed. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 The figure is a schematic diagram of the digital twin platform architecture of the application;

[0043] Figure 2 The figure is a viscosity-temperature curve graph of SAE5W-30 under ideal conditions in Example 1;

[0044] Figure 3 The figure is a viscosity-temperature curve graph of SAE5W-30 under fixed shear rate in Example 1. DETAILED DESCRIPTION

[0045] In order to better explain the application and facilitate understanding, the application is described in detail below in combination with the drawings through specific embodiments.

[0046] The application provides a lubricating oil viscosity monitoring method and system based on parameter coupling and digital twinning, wherein the lubricating oil viscosity monitoring method based on parameter coupling and digital twinning comprises the following steps:

[0047] Step S1, collect the viscosity data of lubricating oil at different temperatures and shear rates using a rheometer, remove outliers, and retain valid data as the model input data for step S2.

[0048] Specifically, data collection is carried out using a TADHR-2 rotary rheometer, which uses a temperature control unit based on Peltier thermoelectric technology, with a temperature regulation range of -40°C to +150°C and a control accuracy of ±0.1°C. The clamp is a 40mm circular aluminum plate, and the sample gap can be adjusted within 100-200μm according to the viscosity characteristics of the oil. During the collection process, the temperature is changed at a rate of 5°C / min within the range of 20°C-80°C, and each temperature point is stable for more than 10 minutes to ensure that the lubricating oil reaches thermal equilibrium. The shear rate is set to 10-38000 s -1 , covering low to high speed mechanical working conditions. For each temperature-shear rate combination, 30 seconds of viscosity data is continuously collected, with a sampling interval of 1 second, and the average value is taken as the valid data.

[0049] The raw data collected in S1 is preprocessed to remove outliers caused by equipment fluctuations or operation errors, and valid samples are retained.

[0050] Step S2, construct a multi-parameter dynamic coupling viscosity-temperature model as follows:

[0051] μ(T,γ)=a(T)(logγ)³+b(T)(logγ)²+c(T)logγ+d(T);

[0052] where μ is the dynamic viscosity, ; T is the temperature, °C; γ is the shear rate, s -1 ; a(T), b(T), c(T) are the coefficients of the shear rate logarithmic terms (logγ) 3 , (logγ) 2 , logγ, respectively; d(T) is the constant term;

[0053] a(T), b(T), c(T), d(T) are determined as follows:

[0054] Let η 预测 =a(logγ)³+b(logγ)²+c(logγ)+d;

[0055] where η 预测 is the predicted viscosity, γ is the shear rate, and a, b, c, d are four parameters. By adjusting the coefficients a, b, c, d, the curve shape of different grades of lubricating oil can be adapted. For any fixed temperature T, the (shear rate, viscosity) data is executed as follows. The iterative optimization process is used to determine the model parameter set x=[a, b, c, d] at that temperature.

[0056] The optimization objective is to minimize the sum of squares of residuals: S(x) =∑[η 实测 -η 预测 ]², where η 实测 is the actual measured viscosity. An adaptive damping factor λ is added to the Levenberg-Marquardt algorithm, and the optimized L-M algorithm is as follows:

[0057] ;

[0058] where J k is the partial derivative matrix of residuals to parameters at the kth iteration, i.e., the Jacobian matrix; λ k is the adaptive damping factor at the kth iteration, r k is the residual vector at the kth iteration, I is the identity matrix, k is the iteration number, and d k is the increment of the parameter vector to be solved at the kth iteration. The initial value x0 of the parameter x to be solved and the initial value of the adaptive damping factor λ are set, and d k is solved. The i-th row of the Jacobian matrix is defined as , where i refers to the i-th experimental data point (γ i , ηmeasuredJ i ).

[0059] The residual ratio is calculated by the following formula:

[0060] ;

[0061] where: is the residual ratio at the kth iteration, is the minimized sum of squares of residuals at the kth iteration, is the minimized sum of squares of residuals at the kth iteration after d k is corrected;

[0062] λ k is adjusted according to ρ k , the parameter vector x is updated, until the sum of squares of residuals is less than the preset convergence threshold or the maximum iteration number is reached, the optimal parameters [a, b, c, d] at the current temperature T are obtained, the optimal parameter groups corresponding to different temperature points are calculated according to the data of different temperature points, , and the obtained , , , , are taken as new data sets, and the functional relationships a(T), b(T), c(T), and d(T) of the parameters a, b, c, and d with temperature T are determined by polynomial fitting.

[0063] Specifically, based on ρ k Adjust λ k :

[0064] If 0.75 < ρ k <1, adjust λ k+1 =λ k / 10, the residual is significantly reduced, the damping factor is reduced, and the convergence is accelerated.

[0065] If 0 < ρ k <0.25, adjust λ k+1 =10λ k If the reduction in residual is insufficient or abnormally large, increase the damping factor to enhance stability.

[0066] If 0.25≤ρ k If ≤0.75, then λ k+1 =λ k , where ρ k Let λ be the residual ratio at the k-th iteration. k Let λ be the adaptive damping factor in the k-th iteration. k+1 This is the adaptive damping factor at the (k+1)th iteration.

[0067] If ρ k If the value is greater than 0, then update parameter x. k+1 =x k +d k , where ρ k x is the residual ratio at the k-th iteration. k+1 Let x be the parameter vector at the (k+1)th iteration. k Let d be the parameter vector at the k-th iteration. k The parameter vector to be determined in the k-th iteration The increment.

[0068] Repeat step S2 until the sum of squared residuals is less than the preset convergence threshold or the maximum number of iterations is reached. Output the optimal parameters [a, b, c, d] at the current temperature T, and calculate multiple sets of different temperature points T based on the data at different temperature points. i The corresponding optimal parameter set (a) i b i c i d i ), will obtain (T i a i ), (T i b i ), (T i c i ), (T i d i) as a new data set, determine the functional relationship a(T), b(T), c(T), d(T) of the parameters a, b, c, d varying with temperature T by polynomial fitting.

[0069] Substitute the fitted functions a(T), b(T), c(T), d(T) into the model expression η 预测 =a(logγ)³+b(logγ)²+c(logγ)+d, to obtain the global viscosity-temperature model μ(T, γ)=a(T)(logγ)³+b(T)(logγ)²+c(T)logγ+d(T), where μ is the dynamic viscosity, , T is the temperature, ℃, γ is the shear rate, s -1 , a(T), b(T), c(T) are the coefficients of the shear rate logarithmic term (logγ) 3 , (logγ) 2 , and logγ, used to express the nonlinear influence of shear rate on viscosity; d(T) is a constant term, reflecting the basic viscosity level under the action of temperature alone; the coefficients a, b, c, d at different temperatures are determined by multiple iteration calculations to ensure that the model fitting goodness R²≥0.999, accurately reflecting the comprehensive influence of temperature and shear rate on viscosity.

[0070] Step S3, build a digital twin platform:

[0071] The digital twin platform is composed of a physical layer, a virtual layer and a data interaction layer, as shown in Figure 1 The physical layer is the temperature sensor, torque sensor and data acquisition card set by the rheometer, which collects the temperature, shear rate-viscosity data of the lubricating oil in real time, transmits it to the edge computing gateway through the industrial bus for filtering and normalization preprocessing, and then uploads it; specifically, the temperature sensor has an accuracy of ±0.1℃; the torque sensor is used to calculate the shear rate, with an accuracy of ±0.5%; the sampling rate of the data acquisition card is ≥1kHz; the virtual layer is based on the multi-parameter dynamic coupling viscosity-temperature model constructed in step S2, and a virtual simulation model is built, and an oil molecule motion sub-model under shear is integrated, which is used to refine the simulation of the micro-molecular behavior of the lubricating oil under the action of temperature and shear, and to assist in improving the accuracy and physical consistency of the viscosity-temperature characteristic simulation. By simulating the change of molecular arrangement with temperature and shear rate, the simulation precision is improved. The data interaction layer uses the OPC UA protocol to realize real-time data synchronization between the physical layer and the virtual layer, with a transmission period ≤100ms, and builds a real-time database to store historical data, supporting data backtracking and model iteration.

[0072] Step S4, based on the model output curve:

[0073] The type of the output curve is determined as a "viscosity-temperature curve" under a fixed shear rate, and in the process of outputting the curve, the temperature T parameter range is selected as -40 DEG C ~ + 150 DEG C, and the shear rate gamma parameter range is selected as 10 ~ 38000 s -1 ; for each temperature point, the corresponding coefficients are fitted by the optimized L-M algorithm, the coefficients are substituted into the model, and the corresponding dynamic viscosity mu is calculated in combination with the selected shear rate; for example, when the shear rate gamma 0 is fixed, the mu corresponding to different temperatures T is calculated, the viscosity curve at the shear rate is obtained, and the curve is displayed through the digital twin platform in step S3.

[0074] Step S5, dynamic correction and performance prediction:

[0075] The digital twin platform compares the physical layer measured viscosity mu 实测 with the virtual layer predicted viscosity mu 预测 , and calculates the relative error:

[0076] ;

[0077] An error threshold is set, when the error exceeds the set threshold, a feedback mechanism is triggered, the optimized L-M algorithm is called to re-fit the coefficients a, b, c, d to correct the model, until the error is not greater than the set threshold, based on the corrected model, the viscosity-temperature curve under different temperature and shear rate combinations is output, specifically, when the error exceeds 5%, the feedback mechanism is triggered, the optimized L-M algorithm is called to re-fit the coefficients a, b, c, d, until the error is less than or equal to 5%, the model is dynamically updated; based on the corrected model, the viscosity-temperature curve under different temperature and shear rate combinations is output, which directly quantifies the viscosity change rule of the lubricating oil under different temperature and shear rate, and provides visual basis for performance evaluation, working condition optimization and failure warning.

[0078] Generate multi-dimensional warning information: when the viscosity change rate is greater than 10% / min, trigger the "viscosity abnormal fluctuation warning"; when the temperature T exceeds the effective lubrication temperature range of the lubricating oil, issue a "critical temperature threshold warning", and further realize the lubrication failure risk assessment, and provide decision support for equipment maintenance.

[0079] The application also provides a lubricating oil viscosity monitoring system based on parameter coupling and digital twinning, which comprises a memory, a processor, and a computer program stored in the memory and running on the processor, and the processor implements the steps of the lubricating oil viscosity monitoring method based on parameter coupling and digital twinning when executing the computer program.

[0080] Embodiment:

[0081] The representative lubricating oil sample SAE5W-30 full synthetic engine oil was selected, and TADHR-2 rotary rheometer, PT100 temperature sensor, high-precision torque sensor and 2 kHz sampling rate data acquisition card were used for data acquisition; the temperature control range was 20-120 ℃, the heating rate was 5 ℃ / min, each temperature point was stabilized for 15 min to ensure thermal equilibrium, the clamp was a 40 mm circular aluminum plate, and the sample loading gap was 150 μm; the shear rate covered 10-38000 s -1 For engine cold start, normal driving and high speed and high load conditions, 30 s of data was continuously collected for each temperature-shear rate combination and the average value was taken, and the data repeatability error was ≤1.5% after 3 times of repeated verification, and part of the key data showed that at 20 ℃, 10 s -1 Shear rate viscosity was 385.2 , 38000 s -1 8.6 , 120 ℃, the corresponding shear rate viscosity was 18.3 and 1.9 .

[0082] A cubic polynomial model was used to describe the relationship between viscosity and temperature and shear rate, and the model expression was μ= a(T)(logγ)³+b(T)(logγ)²+c(T)logγ+d(T), and the parameters [a(T), b(T), c(T), d(T)] were optimized by improved Levenberg-Marquardt algorithm; the initial damping factor λ0=0.01 was set, and the damping factor was adjusted according to the residual ratio ρ k to balance the convergence speed and stability, and after 120 iterations, the model converged, the goodness of fit R²=0.9996, and the parameter fitting results at different temperatures showed that at 20 ℃, a=0.021, b=-0.352, c=2.105, d=5.982, and the parameters changed regularly with the increase of temperature at 60 ℃ and 100 ℃.

[0083] The viscosity-temperature curve at a fixed shear rate was generated, as shown in Figure 3 At idle speed condition (γ=100 s -1 ), the viscosity decreased from 125 at 20 ℃ to 15.2 at 100 ℃, and the temperature sensitivity was significant; at high speed condition (γ=10000 s -1 ), the viscosity decreased from 28.6 at 20 ℃ to 6.8 , the shear thinning effect is obvious; analysis shows that the viscosity of SAE 5W-30 lubricating oil is more sensitive to temperature in the low temperature zone (<40℃), changes slowly in the high temperature zone (>80℃), and the viscosity decreases nonlinearly with the increase of shear rate, which is consistent with the characteristics of pseudoplastic fluid.

[0084] The measured viscosity-temperature curve is compared with the predicted ideal viscosity-temperature curve in real time, and the predicted ideal viscosity-temperature curve is as shown in Figure 2 When the error is greater than 5%, the correction is triggered, for example, at 80℃, γ=5000s -1 Under the working condition, the initial prediction error is 6.5%, and after correction, it is reduced to 1.1%; in terms of performance warning, when the viscosity change rate is greater than 10% / min, the viscosity abnormal fluctuation warning is triggered, and when the temperature is greater than 150℃, the critical temperature warning is issued.

[0085] Although the embodiments of the present application have been shown and described above, it should be understood that the above embodiments are exemplary and should not be construed as limiting the present application, and any modification, modification, replacement and modification of the above embodiments by those skilled in the art shall fall within the scope of the present application.

Claims

1. A method for monitoring lubricating oil viscosity based on parameter coupling and digital twin, characterized in that, Includes the following steps: Step S1: Use a rheometer to collect viscosity data of lubricating oil at different temperatures and shear rates, remove outliers, and retain valid data as model input data for step S2. Step S2: Construct a multi-parameter dynamically coupled viscosity-temperature model as follows: μ(T,γ)=a(T)(logγ)³+b(T)(logγ)²+c(T)logγ+d(T); Where μ is the dynamic viscosity, T is temperature, °C; γ is shear rate, s. -1 a(T), b(T), and c(T) are the logarithmic terms of the shear rate (logγ). 3 (logγ) 2 The coefficients of logγ; d(T) is a constant term; a(T), b(T), c(T), and d(T) are determined as follows: setting 预测 =a(logγ)³+b(logγ)²+c(logγ)+d; Where, η 预测 To predict viscosity, γ is the shear rate, and a, b, c, and d are four parameters. By adding an adaptive damping factor λ to the Levenberg-Marquardt algorithm, the optimized LM algorithm is obtained as follows: ； Among them, J k λ is the partial derivative matrix of the residual with respect to the parameters at the k-th iteration, i.e., the Jacobian matrix; k Let r be the adaptive damping factor in the k-th iteration. k Let I be the residual vector at the k-th iteration, I be the identity matrix, k be the iteration number, and d be the residual vector at the k-th iteration. k The parameter vector to be determined in the k-th iteration The increment; The optimization objective is to minimize the sum of squared residuals: S(x) = Σ[η 实测 -η 预测 ]², where η 实测 This is the actual measured viscosity; The residual ratio is calculated using the following formula: ； in: The residual ratio at the k-th iteration. To minimize the sum of squared residuals in the k-th iteration, For d k Minimize the sum of squared residuals in the k-th iteration of the correction; Based on ρ k Adjust λ k Update the parameter vector x until the sum of squared residuals is less than the preset convergence threshold or the maximum number of iterations is reached, to obtain the optimal parameters [a, b, c, d] at the current temperature T. Calculate the parameters at different temperature points based on the data at different temperature points. The corresponding optimal parameter set , will get , , , As a new dataset, the functional relationships a(T), b(T), c(T), d(T) of parameters a, b, c, d as a function of temperature T were determined by polynomial fitting. Step S3: Build a digital twin platform: The digital twin platform consists of a physical layer, a virtual layer, and a data interaction layer. The physical layer consists of temperature sensors, torque sensors, and a data acquisition card installed on the rheometer to collect real-time temperature and shear rate-viscosity data of the lubricating oil. The virtual layer is based on the multi-parameter dynamic coupling viscosity-temperature model constructed in step S2 to build a virtual simulation model and integrates a sub-model of oil molecule motion under shear. The data interaction layer uses the OPC UA protocol to achieve real-time data synchronization between the physical layer and the virtual layer. Step S4: Based on the model output curve: The output curve is determined to be a "viscosity-temperature curve" under a fixed shear rate. Based on the multi-parameter dynamic coupling viscosity-temperature model in step S2, the corresponding dynamic viscosity μ is calculated in combination with the selected shear rate to obtain the viscosity-temperature curve under the shear rate, which is then displayed through the digital twin platform in step S3. Step S5, Dynamic Correction and Performance Prediction: The digital twin platform compares the measured viscosity μ of the physical layer in real time. 实测 With virtual layer predicted viscosity μ 预测 Calculate the relative error: ； Set an error threshold. When the error exceeds the set threshold, trigger the feedback mechanism and call the optimized LM algorithm to refit the coefficients a, b, c, and d to correct the model until the error is no greater than the set threshold. Based on the corrected model, output the viscosity-temperature curves under different combinations of temperature and shear rate. Generate multi-dimensional early warning information: when the viscosity change rate is >10% / min, trigger "viscosity abnormal fluctuation warning"; when the temperature T exceeds the effective lubrication temperature range of the lubricating oil, issue "critical temperature threshold warning".

2. The lubricating oil viscosity monitoring method based on parameter coupling and digital twin according to claim 1, characterized in that: In step S1, the temperature adjustment range of the rheometer covers -40℃ to +150℃; during the data acquisition process, the temperature is gradually changed within the range of 20℃ to 80℃ at a rate of 5℃ / min, and each temperature point is stabilized for more than 10 minutes, with the shear rate set to 10~38000s. -1 For each temperature-shear rate combination, viscosity data was continuously collected for 30 seconds with a sampling interval of 1 second, and the average value was taken as the valid data.

3. The lubricating oil viscosity monitoring method based on parameter coupling and digital twin according to claim 1, characterized in that: In step S2, based on ρ k Adjust λ k If 0.75 < ρ k <1, adjust λ k+1 =λ k / 10; if 0 < ρ k <0.25, adjust λ k+1 =10λ k If 0.25≤ρ k If ≤0.75, then λ k+1 =λ k ;where ρ k Let λ be the residual ratio at the k-th iteration. k Let λ be the adaptive damping factor in the k-th iteration. k+1 This is the adaptive damping factor at the (k+1)th iteration.

4. The lubricating oil viscosity monitoring method based on parameter coupling and digital twin according to claim 1, characterized in that: In step S2, if ρ k If the value is greater than 0, then update parameter x. k+1 =x k +d k , where ρ k x is the residual ratio at the k-th iteration. k+1 Let x be the parameter vector at the (k+1)th iteration. k Let d be the parameter vector at the k-th iteration. k The parameter vector to be determined in the k-th iteration The increment.

5. The lubricating oil viscosity monitoring method based on parameter coupling and digital twin according to claim 1, characterized in that: In step S3, the data collected by the physical layer of the digital twin platform is transmitted to the edge computing gateway via the industrial bus for filtering, normalization preprocessing, and then uploaded.

6. The lubricating oil viscosity monitoring method based on parameter coupling and digital twin according to claim 1, characterized in that: In step S3, the transmission period of the data interaction layer is ≤100ms.

7. The lubricating oil viscosity monitoring method based on parameter coupling and digital twin according to claim 1, characterized in that: In step S3, a real-time database is built on the digital twin platform to store historical data, supporting data backtracking and model iteration.

8. The lubricating oil viscosity monitoring method based on parameter coupling and digital twin according to claim 1, characterized in that: In step S4, during the output curve process, the temperature parameter T is selected to be in the range of -40℃ to +150℃, and the shear rate parameter γ is selected to be in the range of 10 to 38000 s. -1 .

9. The lubricating oil viscosity monitoring method based on parameter coupling and digital twin according to claim 1, characterized in that: When the error exceeds 5%, a feedback mechanism is triggered, and the optimized LM algorithm is called to refit the coefficients a, b, c, and d until the error is ≤5%, thereby achieving dynamic model updates.

10. A lubricating oil viscosity monitoring system based on parameter coupling and digital twin, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the lubricating oil viscosity monitoring method based on parameter coupling and digital twin as described in any one of claims 1-9.

Images