A method for constructing a global temperature field of a wet clutch with wireless rotary temperature measurement and correction

Abstract

The application discloses a kind of fusion wireless rotary temperature measurement correction's wet clutch global temperature field's construction method, comprising: constructing three-dimensional transient heat conduction model and solving, obtain corresponding working condition under steel sheet global candidate simulation temperature field;Obtain circumferential-radial distribution type wireless temperature measurement data on target steel sheet, extract average temperature rise, radial temperature gradient, circumferential thermal inhomogeneity, heat concentration position and temperature rise rate and other multidimensional thermal characteristics;Based on measured thermal characteristics, multidimensional consistency evaluation of candidate simulation temperature field is carried out on the temperature of measuring point, radial gradient, circumferential distribution and time sequence temperature rise;When consistency does not satisfy the requirement, according to the leading deviation type, model parameters such as friction coefficient, heat transfer coefficient and circumferential contact pressure correction coefficient are corrected in a targeted manner, to generate a reliable temperature field sample.The application upgrades limited measuring point data to global thermal state constraint, realizes the accurate reproduction and intelligent rapid prediction of wet clutch non-uniform temperature field.

Description

Technical Field

[0001] This invention relates to the fields of high-end equipment manufacturing, core transmission components for new energy vehicles, and intelligent manufacturing thermal state monitoring technology, and in particular to a method for constructing a full-domain temperature field for a wet clutch that integrates wireless rotary temperature measurement correction. Background Technology

[0002] Wet clutches are core components of hybrid transmission systems in new energy vehicles and high-precision industrial transmission equipment. Their thermal stability and operational reliability during slip-friction engagement are crucial factors determining the overall transmission efficiency, power stability, and service life of the machine. Under typical high-speed rotating slip-friction engagement conditions, wet clutches continuously generate a large amount of frictional heat at the friction pair interface, causing rapid and drastic temperature rises between the steel plates and friction surfaces. Influenced by multiple complex factors, including assembly clearance errors, uneven local micro-contact states at the friction pair interface, and differences in radial and circumferential heat transfer conditions of the cooling oil, the temperature field of the friction pair is not uniformly distributed. It exhibits not only a significant radial temperature gradient but also a high risk of severe circumferential thermal unevenness and localized heat concentration. This asymmetric temperature rise is a core cause of irreversible steel plate warping, high-temperature thermal degradation of friction materials, localized interface ablation, and transmission vibration failure. For a long time, this has severely restricted the performance improvement, operational adaptability upgrades, and high-reliability industrial applications of high-end wet transmission components, representing a critical technical challenge that urgently needs to be overcome in the high-end transmission field.

[0003] Currently, the mainstream wired temperature measurement solutions in the industry mainly rely on thermocouples and lead wire structures to collect local temperature data of the friction pair. However, wet clutches operate under harsh conditions, constantly in an environment of high-speed rotation, complete oil immersion, and extremely limited installation space. Wired temperature measurement structures have many inherent drawbacks, including difficult wiring, easy wire tangling and fatigue breakage under high-speed conditions, and interference with the equipment's sealing structure. This makes it impossible to achieve long-term stable temperature measurement and fails to meet the core requirement of multi-point synchronous full-area temperature measurement under clutch rotational engagement conditions.

[0004] To address the industry challenge of extracting temperature signals from rotating components, wireless temperature measurement technology is gradually being applied to clutch temperature acquisition. Existing typical technical solutions can be found in CN120741002A and the academic paper "Wireless Temperature Measurement Method for Rotating Friction Elements in Transmission Devices," etc. This type of wireless temperature measurement only completes the hardware architecture, electromagnetic induction power supply, and near-field communication acquisition system development for wet clutch rotating friction pairs. It can only acquire the time-series temperature of discrete measurement points on the inner / middle / outer diameter of the steel sheet, and only verify the accuracy of the measurement point temperature acquisition. It cannot fully reproduce the global temperature distribution characteristics of the clutch friction pair, making it difficult to accurately quantify the actual degree of circumferential thermal non-uniformity, nor can it locate the core area of ​​localized heat concentration. Therefore, it cannot support refined research on the thermal failure mechanism of the friction pair and equipment performance optimization.

[0005] Due to limitations in equipment installation space, operating conditions, and the constraints of the actual measurement environment, experimental temperature measurement methods cannot cover the entire friction pair area. Therefore, temperature field simulation calculation has become the core method for obtaining the complete and continuous temperature distribution of the clutch. The existing simulation method CN121328376A mainly relies on heat conduction theory and finite element model to iteratively calculate multi-physics processes such as frictional heat generation, solid heat conduction, and oil convection heat transfer to solve the temperature field distribution between the steel plate and the friction plate. However, the accuracy of the simulation calculation results is highly dependent on key parameters such as the friction coefficient, interfacial contact pressure, convection heat transfer coefficient, heat distribution coefficient, and boundary conditions. Without the constraints and correction of real temperature measurement data, the simulation model is prone to systematic deviations and cannot accurately reproduce the circumferential non-uniform temperature rise phenomenon caused by local contact differences. The simulation results do not match the actual operating conditions well.

[0006] With the deep integration of intelligent algorithms and industrial simulation technology, data-driven methods such as machine learning and neural networks are gradually being applied to the field of temperature prediction for transmission components, effectively improving the overall efficiency of temperature field calculation. However, current mainstream data-driven prediction models still have significant technical shortcomings. Most models only use basic operating condition parameters such as rotational speed, pressure, inertia, and duration as input variables, and the training samples are mostly derived from pure simulation data that has not been measured and corrected. This not only inherits the inherent systematic errors of traditional simulation models, but also lacks key thermal characteristic parameters such as circumferential thermal non-uniformity, local heat concentration locations, and radial temperature gradients. As a result, the models cannot accurately capture the asymmetric temperature rise characteristics under complex operating conditions, making it difficult to achieve high-precision prediction of the entire temperature field under the rotating engagement condition of wet clutches.

[0007] In summary, the industry lacks an integrated solution for constructing the clutch temperature field that combines rotating operating conditions, actual measurement verification, parameter iterative correction, and intelligent prediction. Currently, the three technologies of wireless temperature measurement hardware, pure finite element simulation, and data-driven prediction are fragmented and cannot simultaneously ensure the authenticity of temperature data, the completeness of temperature field distribution, and the efficiency of prediction. This greatly restricts the technological iteration, performance optimization, and localization upgrade of core components of high-end transmission equipment. Summary of the Invention

[0008] To address the technical problems existing in the prior art, this invention proposes a method for constructing the global temperature field of a wet clutch that integrates wireless rotary temperature measurement correction. Based on multiphysics simulation calculations, the method combines circumferential-radial distributed wireless temperature measurement data to complete simulation verification. By iteratively optimizing model parameters through multidimensional temperature error indices, a high-reliability temperature field sample library and a rapid prediction model are constructed, thus overcoming the shortcomings of existing technologies and meeting the high-precision thermal state monitoring and performance optimization requirements of high-end intelligent transmission equipment.

[0009] On the one hand, to achieve the above objectives, this invention provides a method for constructing the global temperature field of a wet clutch that integrates wireless rotary temperature measurement correction, comprising: A three-dimensional transient heat conduction model considering the circumferential contact pressure difference was constructed and solved to obtain the candidate simulation temperature field of the steel sheet under the corresponding working condition. Under the combined working condition corresponding to the candidate simulated temperature field, temperature sensors are arranged at several circumferential channels and the inner diameter, middle diameter and outer diameter positions of the target steel sheet to obtain the measured temperature at each temperature measurement coordinate at the sampling time. Based on the measured temperature, the average temperature rise, radial temperature gradient, circumferential thermal non-uniformity coefficient, heat concentration position and temperature rise rate are calculated to construct a multi-dimensional measured thermal feature dataset. Based on the actual installation coordinates and sampling time of each temperature sensor, the simulated temperature at the same location and time is extracted from the candidate simulated temperature field. A comprehensive error evaluation function is constructed, which includes the average relative error of the measurement point, radial gradient error, circumferential thermal non-uniformity error and time-series temperature rise error. The consistency evaluation of the candidate simulated temperature field is then performed to determine whether the comprehensive error is less than a preset threshold. When the overall error is greater than or equal to a preset threshold, the dominant deviation type is identified based on the sub-errors, and the parameters to be corrected and the correction direction are determined from the set of parameters to be corrected, which includes the friction coefficient, heat distribution coefficient, heat transfer coefficient and circumferential contact pressure correction coefficient, according to the dominant deviation type. The parameters to be corrected are iteratively updated, and the updated parameters are substituted back into the three-dimensional transient heat conduction model for solution, until the overall error of the corrected simulated temperature field is less than the preset threshold, and the corrected temperature field is obtained. The corrected temperature field that meets the consistency requirements and the corresponding bonding condition parameters are saved as a reliable temperature field sample. The bonding condition parameters are used as input and the circumferential-radial temperature field matrix is ​​used as the output label to train a fast temperature field prediction model, which is used to output the corresponding global temperature field prediction result based on the target bonding condition parameters.

[0010] Preferably, constructing the multidimensional measured thermal feature dataset includes: The first k The temperatures at the inner diameter, middle diameter, and outer diameter measuring points in the circumferential channel are denoted as follows: T ( j , k , t ),in, k =1,2,…,M, representing different data collection groups; j =1(ID), 2(MD), 3(OD) represent sampling points in different radial directions; Calculate the first k Average temperature rise of the circumferential channel for: ; In the formula, The initial temperature before the start of the bonding process; The time of data collection; Calculate the first k Radial temperature gradient of the circumferential channel for: ; In the formula, OD and ID are the radial positions of the outer diameter measuring point and the inner diameter measuring point, respectively; Calculate the first k Circumferential thermal non-uniformity coefficient of the circumferential channel for: ; In the formula, max Indicates the maximum value. minutes This represents the minimum value. mean This represents the average value.

[0011] Preferably, the comprehensive error evaluation function is: ; In the formula, This is the comprehensive error evaluation function. , , and These are the average relative error, radial gradient error, circumferential thermal non-uniformity error, and time-series temperature rise error, respectively. , , and These are the corresponding weighting coefficients.

[0012] Preferably, the preset threshold is determined by calibration of a reference engagement condition that meets repeatability requirements.

[0013] Preferably, the set of parameters to be corrected is: ,in, m The coefficient of friction, or The frictional heat distribution coefficient, The heat transfer coefficient is the inner diameter. The heat transfer coefficient is the outer diameter. To consider the overall heat transfer coefficient, For the first k A contact pressure correction coefficient for each circumferential region is used to make a finite correction to the non-uniform pressure field obtained from the contact analysis.

[0014] Preferably, the step of identifying the dominant deviation type and determining the parameter to be corrected based on the component error includes: Calculate the weighted error terms of the overall error based on the measuring point temperature error, radial temperature gradient error, circumferential thermal non-uniformity error, and temperature rise rate error, respectively. , , and The error type corresponding to the weighted error term with the largest value is determined as the dominant bias type. , , and These are the average relative error, radial gradient error, circumferential thermal non-uniformity error, and time-series temperature rise error, respectively. when When the value is the maximum value among the weighted error terms, the friction coefficient μ and the frictional heat distribution coefficient are corrected first. or ; when When the value is the maximum value among the weighted error terms, the inner diameter heat transfer coefficient should be corrected first. h i and outer diameter heat transfer coefficient h o ; when When the value is the maximum value in the weighted error term, the contact pressure correction coefficient for the corresponding circumferential region is adjusted based on the measured heat concentration location. Make targeted adjustments; when When the value is the maximum value among the weighted error terms, the friction coefficient is corrected first. m Heat distribution coefficient or and overall heat transfer coefficient h r ; The temperature field was resimulated using the corrected model parameters, and the overall error was calculated; when the overall error... ≥Preset threshold Then, based on the updated weighted error terms, the dominant bias type is re-determined, and the corresponding model parameters are iteratively corrected until the comprehensive error is reached. < Output the corrected simulated temperature field.

[0015] Preferably, the training loss function of the rapid temperature field prediction model includes temperature field numerical error, radial temperature gradient error, and circumferential temperature distribution error. The temperature field numerical error is used to constrain the overall temperature difference between the predicted temperature field and the corrected temperature field. The radial temperature gradient error is used to constrain the temperature change of the predicted temperature field along the radius of the steel sheet. The circumferential temperature distribution error is used to constrain the temperature difference and local heat concentration distribution of the predicted temperature field along the circumferential direction.

[0016] Preferably, the training loss function of the rapid temperature field prediction model is as follows: ; In the formula, To train the loss function, For numerical error in the temperature field, For radial temperature gradient error, For circumferential temperature distribution error, , These are the weighting coefficients for the radial temperature gradient error term and the circumferential temperature distribution error term, respectively.

[0017] On the other hand, to achieve the above objectives, the present invention also provides a system for constructing the global temperature field of a wet clutch by integrating wireless rotary temperature measurement correction, comprising: The simulation modeling module is used to construct and solve a three-dimensional transient heat conduction model that considers the differences in circumferential contact pressure, and obtain the candidate simulation temperature field of the steel sheet in the corresponding working condition. The wireless temperature measurement and thermal feature extraction module, under the combined working condition corresponding to the candidate simulated temperature field, arranges temperature sensors at several circumferential channels and the inner diameter, middle diameter, and outer diameter positions of the target steel sheet to obtain the measured temperature at each temperature measurement coordinate at the sampling time, and calculates the average temperature rise, radial temperature gradient, circumferential thermal non-uniformity coefficient, heat concentration position and temperature rise rate based on the measured temperature to construct a multi-dimensional measured thermal feature dataset. The consistency evaluation module is used to extract the simulation temperature at the same location and time in the candidate simulation temperature field based on the actual installation coordinates and sampling time of each temperature sensor, construct a comprehensive error evaluation function that includes the average relative error of the measurement point, radial gradient error, circumferential thermal non-uniformity error and time-series temperature rise error, and perform consistency evaluation on the candidate simulation temperature field to determine whether the comprehensive error is less than a preset threshold. An adaptive parameter correction module is used to identify the dominant deviation type based on the sub-item error when the comprehensive error is greater than or equal to a preset threshold, and determine the parameter to be corrected and the correction direction from the set of parameters to be corrected, which includes friction coefficient, heat distribution coefficient, heat transfer coefficient and circumferential contact pressure correction coefficient, according to the dominant deviation type. The parameter to be corrected is iteratively updated, and the updated parameter is resubmitted into the three-dimensional transient heat conduction model for solution, until the comprehensive error of the corrected simulation temperature field is less than the preset threshold, and the corrected temperature field is obtained. The prediction model construction and output module is used to save the corrected temperature field that meets the consistency requirements and the corresponding joint condition parameters as a reliable temperature field sample. With the joint condition parameters as input and the circumferential-radial temperature field matrix as output label, the module trains a fast temperature field prediction model, which is used to output the corresponding global temperature field prediction result based on the target joint condition parameters.

[0018] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for constructing the global temperature field of a wet clutch with integrated wireless rotary temperature measurement correction.

[0019] Compared with the prior art, the present invention has the following advantages and technical effects: (1) Multidimensional thermal feature quantification: not limited to single-point temperature acquisition, but transforms the measurement point data into standardized thermal features that characterize the temperature distribution of the whole domain, so as to realize the complete characterization of the global thermal state by a small number of measurement points; (2) Multidimensional comprehensive error evaluation: Simultaneously assess temperature amplitude, radial gradient, circumferential non-uniformity, and temperature rise rate to accurately evaluate the simulation's ability to reproduce asymmetric heat concentration; (3) Error-oriented adaptive parameter correction: The heat transfer, friction, and contact pressure parameters are corrected according to the source of the deviation, and the deviation of the simulation system is automatically eliminated without the need for repeated manual parameter adjustment; (4) Intelligent and rapid prediction of global temperature field: Based on the training of surrogate model with real measured and calibrated credible samples, the instantaneous output of global temperature under any working condition is realized, solving the problem of excessive time consumption of finite element simulation and rapid prediction driven by credible temperature field samples. Attached Figure Description

[0020] The accompanying drawings, which form part of this application, are used to provide a further understanding of this application. The illustrative embodiments and descriptions of this application are used to explain this application and do not constitute an undue limitation of this application. In the drawings: Figure 1 This is a flowchart illustrating a method for constructing the global temperature field of a wet clutch that integrates wireless rotary temperature measurement correction, according to an embodiment of the present invention. Figure 2 This is a flowchart of the simulation calculation of the temperature field of the wet clutch according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the measured multidimensional thermal characteristics based on preset temperature measurement coordinates according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the simulation temperature field consistency evaluation process at the measured points in this embodiment of the invention. Figure 5 This is a schematic diagram of the adaptive correction process for simulation model parameters based on measured errors, according to an embodiment of the present invention. Figure 6 This is a schematic diagram illustrating the construction and prediction process of a rapid temperature field prediction model based on calibrated temperature field samples, according to an embodiment of the present invention. Detailed Implementation

[0021] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0022] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0023] This embodiment proposes a method for constructing the global temperature field of a wet clutch that integrates wireless rotary temperature measurement correction, including: A three-dimensional transient heat conduction model considering the circumferential contact pressure difference was constructed and solved to obtain the candidate simulation temperature field of the steel sheet under the corresponding working condition. Under the combined working condition corresponding to the candidate simulated temperature field, temperature sensors are arranged at several circumferential channels and the inner diameter, middle diameter and outer diameter positions of the target steel sheet to obtain the measured temperature at each temperature measurement coordinate at the sampling time. Based on the measured temperature, the average temperature rise, radial temperature gradient, circumferential thermal non-uniformity coefficient, heat concentration position and temperature rise rate are calculated to construct a multi-dimensional measured thermal feature dataset. Based on the actual installation coordinates and sampling time of each temperature sensor, the simulated temperature at the same location and time is extracted from the candidate simulated temperature field. A comprehensive error evaluation function is constructed, which includes the average relative error of the measurement point, radial gradient error, circumferential thermal non-uniformity error and time-series temperature rise error. The consistency evaluation of the candidate simulated temperature field is then performed to determine whether the comprehensive error is less than a preset threshold. When the overall error is greater than or equal to a preset threshold, the dominant deviation type is identified based on the sub-errors, and the parameters to be corrected and the correction direction are determined from the set of parameters to be corrected, which includes the friction coefficient, heat distribution coefficient, heat transfer coefficient and circumferential contact pressure correction coefficient, according to the dominant deviation type. The parameters to be corrected are iteratively updated, and the updated parameters are substituted back into the three-dimensional transient heat conduction model for solution, until the overall error of the corrected simulated temperature field is less than the preset threshold, and the corrected temperature field is obtained. The corrected temperature field that meets the consistency requirements and the corresponding bonding condition parameters are saved as a reliable temperature field sample. The bonding condition parameters are used as input and the circumferential-radial temperature field matrix is ​​used as the output label to train a fast temperature field prediction model, which is used to output the corresponding global temperature field prediction result based on the target bonding condition parameters.

[0024] Specifically, the method in this embodiment transforms the temperature at finite measurement points into multidimensional thermal features that can constrain a complete temperature field. These thermal features are then used to evaluate the consistency of the simulated temperature field and correct its parameters, thereby generating reliable temperature field samples that can be used for training rapid temperature field prediction models. Figure 1 As shown.

[0025] First, based on the clutch structural parameters, material thermophysical properties, engagement conditions, and contact pressure distribution, a candidate transient temperature field model for the rotational slippage process is established. The solution result is denoted as... , This represents the temperature data acquired at all times and locations during the simulation, where r , i , z These are the radial, circumferential, and axial coordinates of the steel sheet, respectively. t This refers to the moment during the joining process.

[0026] Secondly, circumferential-radial distributed wireless temperature measurement channels were arranged on the target steel sheet consistent with the simulation conditions to obtain the measured temperatures at different circumferential channel positions and at the inner diameter, middle diameter, and outer diameter positions. It also calculates the measured thermal characteristics such as average temperature rise, radial temperature gradient, circumferential thermal non-uniformity, location of heat concentration, and rate of temperature rise.

[0027] Then, based on the actual installation coordinates and sampling time of each temperature sensor, the candidate simulated temperature field is... Extract the simulated temperature at the same location and time. , It represents the simulated temperature at the time and location corresponding to the measured temperature, and calculates the temperature error at the measuring point, radial gradient error, circumferential thermal non-uniformity error, thermal concentration location deviation, and temperature rise rate error to form a comprehensive consistency evaluation result.

[0028] When the error between the candidate simulated temperature field and the measured thermal characteristics does not meet the preset requirements, the parameters to be corrected and their correction directions are determined according to different error types. The friction coefficient, heat distribution coefficient, inner and outer diameter heat transfer coefficient, comprehensive heat transfer coefficient and circumferential contact pressure correction coefficient are updated in a directional manner until the corrected complete temperature field meets the requirements of the measured consistency.

[0029] Finally, the complete corrected temperature field that meets the consistency requirements and its corresponding operating parameters are saved as a reliable temperature field sample, and a rapid temperature field prediction model is constructed to realize the rapid prediction of the non-uniform temperature field of the steel sheet from the joint operating parameters.

[0030] Specifically, the following steps are included: S1. Simulation calculation of temperature field of wet clutch considering circumferential thermal non-uniformity: Based on the structural parameters, material thermophysical properties, and engagement conditions of the wet clutch steel plates and friction plates, a contact mechanics model of the friction elements is established to obtain the contact pressure of the friction elements. .

[0031] A transient heat conduction model in a three-dimensional cylindrical coordinate system is established, considering the circumferential contact pressure differences. The model inputs include the rotational speed of the active end. , Engagement hydraulic pressure p (t), moment of inertia of the passive end J , coupling time t And the inner and outer diameters, thickness and material parameters of the clutch.

[0032] Considering the annular structure of the clutch steel plates and the differences in circumferential contact pressure and heat transfer conditions during actual engagement, a three-dimensional transient heat conduction equation is established in cylindrical coordinates: ; In the formula, T For temperature, l、r , c These are the thermal conductivity, density, and specific heat capacity of the friction element, respectively. z、θ and r These represent the axial, circumferential, and radial coordinates in a cylindrical coordinate system, respectively.

[0033] Assuming all frictional work is converted into heat generated by friction, the heat flux density between the friction pairs is: ; In the formula, It is the coefficient of friction with respect to relative angular velocity.

[0034] Relative angular velocity for: ; In the formula, It's about clutch torque. J It is the inertia of the passive component.

[0035] After setting the frictional heat source, contact pressure field, and thermal boundary conditions, the transient heat conduction solution is performed on the above model to obtain the temperature of each calculation node in the steel sheet domain at each sampling time, thus obtaining the complete candidate temperature field of the steel sheet under the corresponding working condition. ,like Figure 2 As shown.

[0036] S2. Construction of measured thermal characteristics based on preset temperature measurement coordinates: For the rotational joining process corresponding to the candidate simulation conditions, multiple circumferential and radial coordinates on the steel sheet are predetermined, and the preset temperature measurement coordinates are obtained at the sampling time. The actual measured temperature at the location.

[0037] Establish a correspondence between the measured temperature data and the temperature measurement coordinates, initial temperature, active end rotation speed, engagement oil pressure, passive end rotational inertia, and engagement time.

[0038] A multidimensional thermal feature quantification algorithm is proposed: based on batch calculation of five standardized thermal feature indices—average temperature rise, radial temperature gradient, circumferential thermal non-uniformity coefficient, heat concentration location, and temperature rise rate—from discrete measurement points, the one-dimensional time-series measurement point data is upgraded into a spatial distribution feature dataset of a constrained global temperature field, as detailed below: This embodiment uses four circumferential channels, i.e., M=4. Each circumferential channel includes temperature measurements in three radial directions: inner diameter ID, middle diameter MD, and outer diameter OD. The first... k The temperatures at the inner diameter, middle diameter, and outer diameter measuring points in the circumferential channel are denoted as follows: T ( j , k , t ),in, k =1,2,…,M, representing different data collection groups; j =1(ID), 2(MD), 3(OD) represent sampling points in different radial directions; Calculate the first k Average temperature rise of the circumferential channel for: ; In the formula, The initial temperature before the start of the bonding process; This refers to the time of data collection.

[0039] Calculate the first... k Radial temperature gradient of the circumferential channel for: ; In the formula, OD and ID are the radial positions of the outer diameter measuring point and the inner diameter measuring point, respectively; The radial temperature gradient is used to characterize the degree of temperature variation along the radial direction at different circumferential locations.

[0040] Calculate the first k Circumferential thermal non-uniformity coefficient of the circumferential channel for: ; In the formula, max Indicates the maximum value. minutes This represents the minimum value. mean This represents the average value.

[0041] The larger the value, the more pronounced the difference in circumferential temperature distribution in the steel sheet.

[0042] The measured temperature at the first sampling time was calculated based on the average temperature rise at adjacent sampling times. k Temperature rise rate of the channel group: ; In the formula, Δ t This represents the time interval between adjacent sampling times.

[0043] Finally, the temperature at each measuring point is obtained. Radial temperature gradient Circumferential thermal non-uniformity coefficient Location of heat concentration k * and rate of temperature rise Hot feature datasets, such as Figure 3 As shown.

[0044] The aforementioned temperature and thermal characteristics at the measurement points are used to determine the corresponding evaluation locations in the simulated temperature field, and to conduct a consistency evaluation of the measurement point temperature, radial gradient, circumferential thermal non-uniformity, and temperature rise rate. They also serve as the experimental basis for simulation parameter correction and the construction of a rapid temperature field prediction model.

[0045] S3. Consistency evaluation of simulated temperature field based on measured points: To evaluate the consistency between the candidate simulated temperature field and the actual engagement process of the wet clutch, temperature field simulation calculations and wireless temperature measurement experiments were conducted under the same conditions of driving end speed, engagement oil pressure, driven end inertia, initial temperature, and engagement time. A four-dimensional weighted comprehensive error evaluation function was constructed, simultaneously constraining the numerical error of the measured temperature, radial gradient error, circumferential thermal unevenness error, and the temporal error of the temperature rise rate, rather than just the temperature deviation of a single measured point. The simulation model's ability to reproduce circumferential local heat concentration, radial temperature difference, and dynamic temperature rise trend was quantitatively determined, forming a standard for judging the suitability of the simulation model.

[0046] Based on the actual installation coordinates of each temperature sensor on the steel sheet, the simulated temperatures at the same circumferential position, radial position, and time are extracted from the candidate simulated temperature field. .

[0047] This step directly compares the temperature at the same location and time in the simulated temperature field.

[0048] First, calculate the average relative error of all measuring points within the evaluation time range. : ; In the formula, This represents the number of sampling times used in the evaluation. Used to evaluate the overall ability of the simulation model to reproduce the temperature values ​​at each measuring point.

[0049] Secondly, the radial gradient error of each circumferential channel is calculated based on the simulated temperature and the measured temperature: ; Simultaneously, the overall circumferential thermal non-uniformity coefficient was calculated from both the simulated and measured temperatures. and .

[0050] The circumferential thermal non-uniformity error is expressed as: ; The simulation results and measured results were compared to determine the location of heat concentration. Whether they are consistent is used to determine whether the simulation model can reproduce the circumferential local heat concentration area.

[0051] Based on the average temperature rise change at adjacent sampling times, the temperature rise rate under simulation and actual measurement conditions is calculated, and the time-series temperature rise error is obtained. .

[0052] ; Construct a comprehensive error evaluation function: ; In the formula, , , and These are the average relative error, radial gradient error, circumferential thermal non-uniformity error, and time-series temperature rise error, respectively. , , and For the corresponding weight coefficients ( α =0.40, β =0.20, c =0.25, d =0.15). Overall threshold The preset threshold is determined by calibration of the benchmark joint condition that meets the repeatability requirements; specifically, it can be taken as the comprehensive error of the benchmark sample. E The 95th percentile or its equivalent statistical upper limit.

[0053] When the comprehensive error corresponding to the candidate simulated temperature field satisfies < When the candidate simulated temperature field is considered to have acceptable consistency with the measured results under the combined working condition in terms of measuring point temperature, radial temperature gradient, circumferential thermal non-uniformity, and temperature rise trend, the corresponding model parameters and complete temperature field are retained as subsequent reliable temperature field samples; when the comprehensive error satisfies At this point, it was determined that the candidate simulated temperature field did not yet meet the requirements for consistency with actual measurements. Therefore, the temperature error at the measurement point was considered... Radial gradient error Circumferential thermal non-uniformity error and temperature rise rate error The error types corresponding to the weighted error terms with the largest values ​​are compared and determined as the dominant deviation type of the current temperature field. Based on this dominant deviation type, the corresponding model parameters to be corrected are selected, and the process proceeds to the subsequent adaptive correction stage of the simulation model parameters. Figure 4 As shown.

[0054] S4. Adaptive calibration of simulation model parameters based on measured errors: When the comprehensive error calculated in S3 E Greater than or equal to the preset threshold This indicates a significant deviation between the candidate simulated temperature field and the measured temperature, necessitating correction of the temperature field simulation model parameters. Therefore, a multi-dimensional set of parameters to be corrected is established, including the friction coefficient, heat distribution coefficient, inner and outer diameter heat transfer coefficients, and circumferential contact pressure correction coefficient. Based on the identification of dominant deviation types according to the sub-errors, the corresponding physical parameters are iteratively corrected in a targeted manner. This closed-loop approach reduces the deviation between the simulated and measured thermal characteristics, automatically generating a reliable temperature field that matches actual operating conditions, forming a closed-loop process of "measurement-evaluation-correction-resimulation." The specific operations are as follows: coefficient of friction m Frictional heat distribution coefficient or Inner diameter heat transfer coefficient h i outer diameter heat transfer coefficient h o Overall heat transfer coefficient h r and contact pressure field correction coefficient As the parameter to be corrected.

[0055] The set of parameters to be corrected is represented as .

[0056] in, For the first k A contact pressure correction factor for each circumferential region is used to perform a finite correction on the non-uniform pressure field obtained from the contact analysis. The corrected contact pressure is expressed as: ; Based on the magnitude of each error in S3, determine the parameters to be corrected first: (1) When the temperature error at the measuring point When the value is the maximum among the four weighted error terms, the friction coefficient is corrected first. m and frictional heat distribution coefficient or This is to adjust the overall level of frictional heat generation.

[0057] (2) When the radial temperature gradient error When the value is the maximum among the four weighted error terms, the inner diameter heat transfer coefficient should be corrected first. h iand outer diameter heat transfer coefficient h o This is to adjust the radial temperature distribution of the steel sheet.

[0058] (3) When the circumferential temperature distribution error When the value is the maximum among the four weighted error terms, the contact pressure correction coefficient for the corresponding circumferential region is determined based on the measured heat concentration location. Targeted adjustments were made to improve the ability of the simulated temperature field to reproduce circumferential thermal inhomogeneities.

[0059] (4) When the temperature rise rate error When the value is the maximum among the four weighted error terms, the friction coefficient is corrected first. m Heat distribution coefficient or and overall heat transfer coefficient h r This is to adjust the temperature change trend over time.

[0060] Based on comprehensive error E The minimum is the correction target, and the parameters to be corrected are iteratively updated within the allowable range of each parameter.

[0061] After the m-th iteration, the updated parameter set will be... X Substitute the transient temperature field model back into the model, recalculate the complete simulated temperature field, and extract the simulated temperature at each sensor location again. Recalculate the average relative error of the measuring point temperature, radial gradient error, circumferential thermal non-uniformity error, temporal temperature rise error, and overall error according to the method described in S3. When the overall error after the m-th iteration satisfies... Then stop parameter correction. Let the parameter set for the m-th iteration be denoted as... The corresponding temperature field Recorded as Both are stored together as a reliable temperature field sample.

[0062] When the number of iterations reaches the preset maximum value and the overall error still does not meet the requirements When required, the parameter calibration process under that operating condition is paused, and the candidate temperature field corresponding to that operating condition is temporarily excluded from the reliable temperature field sample set. Subsequently, the measured temperature data, contact pressure field, and model boundary conditions for that operating condition are checked; after checking and correcting the relevant inputs or boundary conditions, simulation calculations and consistency evaluations are carried out again, such as... Figure 5 As shown.

[0063] S5. Construction of a fast temperature field prediction model based on calibrated temperature field samples: After completing the simulation model parameter calibration, for those that meet the comprehensive error... The required joint conditions will be determined by the corresponding calibration model parameters. and complete temperature field correction They are collectively stored as a reliable temperature field sample.

[0064] A surrogate model is constructed using the corrected high-precision global temperature field as a label and the operating parameters as input. A composite loss function is designed to simultaneously constrain the temperature value, radial gradient, and circumferential distribution. After training, the complete global temperature field can be output instantaneously by inputting the rotational speed, oil pressure, and inertia, eliminating the need for repeated finite element transient iterations and significantly improving computational efficiency.

[0065] Select the axial position corresponding to the temperature sensor mounting surface. Discretize the corrected temperature field at that location into a single point. radial nodes and The circumferential-radial temperature field matrix formed by the circumferential nodes is denoted as: ; in, This indicates the first joint condition. i Each sampling time, This is the global circumferential-radial corrected temperature field matrix of the steel sheet at that moment.

[0066] The model input corresponding to the correction temperature field matrix is ​​the junction parameter under this operating condition, and the input vector is represented as: ; The basic input parameters for the rapid temperature field prediction model include the active end rotation speed. , Engagement hydraulic pressure Moment of inertia at the passive end J , coupling time and initial temperature .

[0067] Therefore, a training sample set for a rapid temperature field prediction model is constructed: ; The model takes the jointing parameters and the jointing time as inputs, and the circumferential-radial temperature field of the steel sheet after being constrained by measured thermal characteristics and corrected by parameters as the output label.

[0068] To avoid inflated model evaluation results due to adjacent temperature data from the same bonding process being included in both the training and testing sets, the samples are divided according to the bonding conditions. All time samples under the same bonding condition are assigned to only one set in the training, validation, or testing sets.

[0069] A rapid temperature field prediction model is established, consisting of a working condition feature encoding module and a temperature field decoding module. The working condition feature encoding module is used to extract the nonlinear correlation features between the active end rotational speed, engagement oil pressure, passive end rotational inertia, engagement time, and initial temperature; the temperature field decoding module is used to map the encoded working condition features into the circumferential-radial temperature field matrix at the corresponding time.

[0070] The model's prediction results are expressed as follows: ; in, For a rapid temperature field prediction model, f These are the parameters to be trained on the model. This is the circumferential-radial temperature field matrix predicted by the model.

[0071] To ensure that the predicted temperature field not only closely approximates the corrected temperature field in terms of temperature values, but also maintains the characteristics of radial temperature variation and circumferential thermal inhomogeneity, the model training loss function is composed of the numerical error of the temperature field, the radial temperature gradient error, and the circumferential temperature distribution error. ; Among them, the numerical error of the temperature field The formula used to constrain the overall temperature difference between the predicted and corrected temperature fields is as follows: ; in a , b It is the node number.

[0072] Radial temperature gradient error The formula used to constrain and predict temperature changes along the radial direction of the steel sheet is: ; Circumferential temperature distribution error The formula used to constrain the predicted temperature difference and local heat concentration distribution along the circumferential direction of the temperature field is as follows: ; Among them, when calculated to season b +1=1 to ensure the continuity of the circumferential boundary of the steel sheet; and These are the weighting coefficients for the radial temperature gradient error term and the circumferential temperature distribution error term, respectively, and their values ​​are determined through the validation set.

[0073] During training, the model parameters are continuously updated using the training set. The model training epochs and loss term weights are determined using the validation set. When the loss function on the validation set reaches its minimum or no longer decreases significantly within a preset number of training epochs, the corresponding model parameters are saved.

[0074] Subsequently, the test set was used to evaluate the model's ability to predict the temperature field under conditions not included in the training. The evaluation indicators included global average temperature error, peak temperature error, radial temperature gradient error, circumferential heat concentration location deviation, and temperature rise trend error.

[0075] After training, given the initial rotational speed of the driving end, engagement oil pressure, moment of inertia of the driven end, engagement time, and initial temperature, the rapid temperature field prediction model can directly output the global circumferential-radial predicted temperature field of the steel sheet at the corresponding time. This enables rapid calculation of the temperature field and identification of heat concentration areas during the rotational engagement process of the wet clutch. The above process is as follows: Figure 6 As shown.

[0076] This embodiment also provides a system for constructing the global temperature field of a wet clutch by integrating wireless rotary temperature measurement correction, including: The simulation modeling module is used to construct and solve a three-dimensional transient heat conduction model that considers the differences in circumferential contact pressure, and obtain the candidate simulation temperature field of the steel sheet in the corresponding working condition. The wireless temperature measurement and thermal feature extraction module, under the combined working condition corresponding to the candidate simulated temperature field, arranges temperature sensors at several circumferential channels and the inner diameter, middle diameter, and outer diameter positions of the target steel sheet to obtain the measured temperature at each temperature measurement coordinate at the sampling time, and calculates the average temperature rise, radial temperature gradient, circumferential thermal non-uniformity coefficient, heat concentration position and temperature rise rate based on the measured temperature to construct a multi-dimensional measured thermal feature dataset. The consistency evaluation module is used to extract the simulation temperature at the same location and time in the candidate simulation temperature field based on the actual installation coordinates and sampling time of each temperature sensor, construct a comprehensive error evaluation function that includes the average relative error of the measurement point, radial gradient error, circumferential thermal non-uniformity error and time-series temperature rise error, and perform consistency evaluation on the candidate simulation temperature field to determine whether the comprehensive error is less than a preset threshold. An adaptive parameter correction module is used to identify the dominant deviation type based on the sub-item error when the comprehensive error is greater than or equal to a preset threshold, and determine the parameter to be corrected and the correction direction from the set of parameters to be corrected, which includes friction coefficient, heat distribution coefficient, heat transfer coefficient and circumferential contact pressure correction coefficient, according to the dominant deviation type. The parameter to be corrected is iteratively updated, and the updated parameter is resubmitted into the three-dimensional transient heat conduction model for solution, until the comprehensive error of the corrected simulation temperature field is less than the preset threshold, and the corrected temperature field is obtained. The prediction model construction and output module is used to save the corrected temperature field that meets the consistency requirements and the corresponding joint condition parameters as a reliable temperature field sample. With the joint condition parameters as input and the circumferential-radial temperature field matrix as output label, the module trains a fast temperature field prediction model, which is used to output the corresponding global temperature field prediction result based on the target joint condition parameters.

[0077] An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, when the processor executes the computer program, it implements a method for constructing a wet clutch global temperature field with fused wireless rotary temperature measurement correction.

[0078] The above are merely preferred embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for constructing the global temperature field of a wet clutch integrating wireless rotary temperature measurement correction, characterized in that, include: A three-dimensional transient heat conduction model considering the circumferential contact pressure difference was constructed and solved to obtain the candidate simulation temperature field of the steel sheet under the corresponding working condition. Under the combined working condition corresponding to the candidate simulated temperature field, temperature sensors are arranged at several circumferential channels and the inner diameter, middle diameter and outer diameter positions of the target steel sheet to obtain the measured temperature at each temperature measurement coordinate at the sampling time. Based on the measured temperature, the average temperature rise, radial temperature gradient, circumferential thermal non-uniformity coefficient, heat concentration position and temperature rise rate are calculated to construct a multi-dimensional measured thermal feature dataset. Based on the actual installation coordinates and sampling time of each temperature sensor, the simulated temperature at the same location and time is extracted from the candidate simulated temperature field. A comprehensive error evaluation function is constructed, which includes the average relative error of the measurement point, radial gradient error, circumferential thermal non-uniformity error and time-series temperature rise error. The consistency evaluation of the candidate simulated temperature field is then performed to determine whether the comprehensive error is less than a preset threshold. When the overall error is greater than or equal to a preset threshold, the dominant deviation type is identified based on the sub-errors, and the parameters to be corrected and the correction direction are determined from the set of parameters to be corrected, which includes the friction coefficient, heat distribution coefficient, heat transfer coefficient and circumferential contact pressure correction coefficient, according to the dominant deviation type. The parameters to be corrected are iteratively updated, and the updated parameters are substituted back into the three-dimensional transient heat conduction model for solution, until the overall error of the corrected simulated temperature field is less than the preset threshold, and the corrected temperature field is obtained. The corrected temperature field that meets the consistency requirements and the corresponding bonding condition parameters are saved as a reliable temperature field sample. The bonding condition parameters are used as input and the circumferential-radial temperature field matrix is ​​used as the output label to train a fast temperature field prediction model, which is used to output the corresponding global temperature field prediction result based on the target bonding condition parameters.

2. The method for constructing the global temperature field of a wet clutch with integrated wireless rotary temperature measurement correction according to claim 1, characterized in that, The multidimensional measured thermal feature dataset is constructed as follows: The first k The temperatures at the inner diameter, middle diameter, and outer diameter measuring points in the circumferential channel are denoted as follows: T ( j , k , t ),in, k =1,2,…,M, representing different data collection groups; j =1(ID), 2(MD), 3(OD) represent sampling points in different radial directions; Calculate the first k Average temperature rise of the circumferential channel for: ； In the formula, The initial temperature before the start of the bonding process; The time of data collection; Calculate the first k Radial temperature gradient of the circumferential channel for: ； In the formula, OD and ID are the radial positions of the outer diameter measuring point and the inner diameter measuring point, respectively; Calculate the first k Circumferential thermal non-uniformity coefficient of the circumferential channel for: ； In the formula, max Indicates the maximum value. min This represents the minimum value. mean This represents the average value.

3. The method for constructing the global temperature field of a wet clutch with integrated wireless rotary temperature measurement correction according to claim 1, characterized in that, The comprehensive error evaluation function is: ； In the formula, This is the comprehensive error evaluation function. , , and These are the average relative error, radial gradient error, circumferential thermal non-uniformity error, and time-series temperature rise error, respectively. , , and These are the corresponding weighting coefficients.

4. The method for constructing the global temperature field of a wet clutch with integrated wireless rotary temperature measurement and correction according to claim 3, characterized in that, The preset threshold is determined by calibration of the benchmark engagement condition that meets the repeatability requirements.

5. The method for constructing the global temperature field of a wet clutch with integrated wireless rotary temperature measurement correction according to claim 1, characterized in that, The set of parameters to be corrected is as follows: ,in, μ The coefficient of friction, η The frictional heat distribution coefficient, The heat transfer coefficient is the inner diameter. The heat transfer coefficient is the outer diameter. To consider the overall heat transfer coefficient, For the first k A contact pressure correction coefficient for each circumferential region is used to make a finite correction to the non-uniform pressure field obtained from the contact analysis.

6. The method for constructing the global temperature field of a wet clutch with integrated wireless rotary temperature measurement correction according to claim 5, characterized in that, The step of identifying the dominant deviation type and determining the parameter to be corrected based on the component error includes: Calculate the weighted error terms of the overall error based on the measuring point temperature error, radial temperature gradient error, circumferential thermal non-uniformity error, and temperature rise rate error, respectively. , , and The error type corresponding to the weighted error term with the largest value is determined as the dominant bias type. , , and These are the average relative error, radial gradient error, circumferential thermal non-uniformity error, and time-series temperature rise error, respectively. when When the value is the maximum value among the weighted error terms, the friction coefficient μ and the frictional heat distribution coefficient are corrected first. η ; when When the value is the maximum value among the weighted error terms, the inner diameter heat transfer coefficient should be corrected first. h i and outer diameter heat transfer coefficient h o ; when When the value is the maximum value in the weighted error term, the contact pressure correction coefficient for the corresponding circumferential region is adjusted based on the measured heat concentration location. Make targeted adjustments; when When the value is the maximum value among the weighted error terms, the friction coefficient is corrected first. μ Heat distribution coefficient η and overall heat transfer coefficient h r ; The temperature field was resimulated using the corrected model parameters, and the overall error was calculated; when the overall error... ≥Preset threshold Then, based on the updated weighted error terms, the dominant bias type is re-determined, and the corresponding model parameters are iteratively corrected until the comprehensive error is reached. < Output the corrected simulated temperature field.

7. The method for constructing the global temperature field of a wet clutch with integrated wireless rotary temperature measurement correction according to claim 1, characterized in that, The training loss function of the rapid temperature field prediction model includes temperature field numerical error, radial temperature gradient error, and circumferential temperature distribution error. The temperature field numerical error is used to constrain the overall temperature difference between the predicted temperature field and the corrected temperature field. The radial temperature gradient error is used to constrain the temperature change of the predicted temperature field along the radius of the steel sheet. The circumferential temperature distribution error is used to constrain the temperature difference and local heat concentration distribution of the predicted temperature field along the circumferential direction.

8. The method for constructing the global temperature field of a wet clutch with integrated wireless rotary temperature measurement correction according to claim 7, characterized in that, The training loss function of the rapid temperature field prediction model is as follows: ； In the formula, To train the loss function, For numerical error in the temperature field, For radial temperature gradient error, For circumferential temperature distribution error, , These are the weighting coefficients for the radial temperature gradient error term and the circumferential temperature distribution error term, respectively.

9. A system for constructing the global temperature field of a wet clutch integrating wireless rotary temperature measurement correction, used to implement the method for constructing the global temperature field of a wet clutch integrating wireless rotary temperature measurement correction as described in any one of claims 1-8, characterized in that, include: The simulation modeling module is used to construct and solve a three-dimensional transient heat conduction model that considers the differences in circumferential contact pressure, and obtain the candidate simulation temperature field of the steel sheet in the corresponding working condition. The wireless temperature measurement and thermal feature extraction module, under the combined working condition corresponding to the candidate simulated temperature field, arranges temperature sensors at several circumferential channels and the inner diameter, middle diameter, and outer diameter positions of the target steel sheet to obtain the measured temperature at each temperature measurement coordinate at the sampling time, and calculates the average temperature rise, radial temperature gradient, circumferential thermal non-uniformity coefficient, heat concentration position and temperature rise rate based on the measured temperature to construct a multi-dimensional measured thermal feature dataset. The consistency evaluation module is used to extract the simulation temperature at the same location and time in the candidate simulation temperature field based on the actual installation coordinates and sampling time of each temperature sensor, construct a comprehensive error evaluation function that includes the average relative error of the measurement point, radial gradient error, circumferential thermal non-uniformity error and time-series temperature rise error, and perform consistency evaluation on the candidate simulation temperature field to determine whether the comprehensive error is less than a preset threshold. An adaptive parameter correction module is used to identify the dominant deviation type based on the sub-item error when the comprehensive error is greater than or equal to a preset threshold, and determine the parameter to be corrected and the correction direction from the set of parameters to be corrected, which includes friction coefficient, heat distribution coefficient, heat transfer coefficient and circumferential contact pressure correction coefficient, according to the dominant deviation type. The parameter to be corrected is iteratively updated, and the updated parameter is resubmitted into the three-dimensional transient heat conduction model for solution, until the comprehensive error of the corrected simulation temperature field is less than the preset threshold, and the corrected temperature field is obtained. The prediction model construction and output module is used to save the corrected temperature field that meets the consistency requirements and the corresponding joint condition parameters as a reliable temperature field sample. With the joint condition parameters as input and the circumferential-radial temperature field matrix as output label, the module trains a fast temperature field prediction model, which is used to output the corresponding global temperature field prediction result based on the target joint condition parameters.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method for constructing the global temperature field of a wet clutch with fusion wireless rotary temperature measurement correction as described in any one of claims 1-8.

Images