Efficient friction element running-in method

Abstract

The present application relates to a kind of high-efficiency friction element running-in method, comprising: the running-in condition of pre-setting is introduced into the clutch test control system of pre-built, the cycle test is carried out to friction sample, and the friction torque data of friction sample in the running-in condition in the preset cycle is extracted;Friction torque data is verified for consistency, if meet, then microtopography consistency evaluation is carried out;After the consistency evaluation of microtopography, the cycle number is recorded, and the average friction coefficient of the last cycle under specific condition is extracted;The friction coefficient estimation and the prediction of the remaining running-in number of the running-in process of the friction element of the same friction material but different size series are carried out.The present application proposes a running-in number prediction method based on machine learning for the friction element of the same material and different size series, without the need for running-in dynamics consistency and actual topography stability discrimination, directly predict the required running-in number by average friction experimental data of early running-in test.

Description

Technical Field

[0001] This invention relates to the field of automotive clutch technology, and in particular to a method for efficient break-in of friction elements. Background Technology

[0002] The clutch is a crucial component in all types of stepped transmission vehicles, and the slip characteristics of its friction elements directly determine the quality of gear shifting. Friction elements undergo a break-in phase, a stable wear phase, and a severe wear phase throughout their lifespan. The break-in phase is critical for achieving ideal slip performance. During break-in, the initially rough surfaces mesh, and the frictional properties gradually stabilize. The quality of the break-in directly determines the material's service life. If friction elements are used directly without break-in, the surface contact of the friction pair will be extremely uneven. The friction elements will bear non-uniform loads throughout their lifespan, resulting in high stress in the actual contact area. When the contact surfaces move relative to each other, adhesive wear or direct surface abrasion occurs, causing heat generation and further wear. This leads to premature damage to mechanical parts, significantly shortening their lifespan and greatly reducing vehicle performance.

[0003] To address these issues, pre-run-in of clutch friction plates is typically required before normal use. During the run-in process, the micro-protrusions on the friction surfaces are gradually sheared and smoothed, increasing the actual contact area between the two surfaces and maintaining optimal surface contact, thus reducing contact stress. The paper "Study on Friction Performance of Copper-Based Powder Metallurgy Dry Friction Pairs During Run-in" presents the run-in performance of friction elements under long-term sliding friction conditions; however, the run-in effect is affected by different operating conditions. Patents describe a clutch friction plate run-in device that reduces manual labor intensity and improves work efficiency; and patents also describe a clutch run-in device and its control circuit and method. However, while there is research and development on run-in devices, no research or related patents describe how to evaluate the changes in friction characteristics during the run-in process. This makes it impossible to determine the end time of the run-in, and there are no indicators to evaluate the run-in effect, failing to effectively guide the run-in test of friction plates. This results in low run-in efficiency and unsatisfactory results, directly affecting the subsequent normal performance of the friction plates. Summary of the Invention

[0004] The purpose of this invention is to propose an efficient break-in method for friction elements to address the problems existing in the prior art. Based on the actual speed and oil pressure characteristics of the clutch during vehicle operation, a cyclic break-in process is established. The consistency of torque changes under adjacent cycles is compared, and then the stability of the actual morphological characteristics of the friction elements is compared to determine the number of break-in cycles and the break-in end time. Based on this, a machine learning model based on the average friction coefficient under cyclic conditions is proposed, which can be used to predict the number of break-in cycles for friction elements of the same material and different size series.

[0005] To achieve the above objectives, the present invention provides the following solution:

[0006] A method for efficient break-in of friction elements includes:

[0007] The preset running-in conditions are imported into the pre-built clutch test control system, and the friction sample is subjected to cyclic test to extract the friction torque data of the friction sample under the running-in conditions in the preset number of cycles.

[0008] The friction torque data is verified for consistency. If the consistency is satisfied, the microstructure consistency is evaluated. If the consistency is not satisfied, the friction sample is subjected to cyclic testing.

[0009] After evaluating the consistency of the microstructure, the final number of cycles was recorded;

[0010] Based on the final number of cycles, the friction elements are run-in; wherein, for the first friction element of the same friction material and the same size series: only the final number of cycles is required for running-in; for the second friction element of the same friction material but a different size series: based on the final number of cycles, the second friction element is run-in, the average friction coefficient is extracted, and based on the average friction coefficient, the friction coefficient of the second friction element during the running-in process and the remaining number of running-in cycles are estimated and predicted.

[0011] Optionally, the friction sample includes: a mating steel sheet and a friction plate;

[0012] Cyclic testing of friction samples includes:

[0013] The friction sample is subjected to n cycles, each cycle including m running-in conditions; wherein the running-in conditions include: first condition: 0.6p, 1200n1, second condition: 1.2p, 1200n1, third condition: 1.8p, 1200n1, fourth condition: 2.4p, 1200n1, fifth condition: 1.2p, 600n1, sixth condition: 1.2p, 1200n1, seventh condition: 1.2p, 1800n1, eighth condition: 1.2p, 2400n1.

[0014] Optionally, verifying the consistency of the friction torque data includes:

[0015] Extract the friction torque data of the friction sample under m running conditions in the (n-2)th and nth cycles;

[0016] The friction torque data is preprocessed; wherein the preprocessing includes: deburring and filtering;

[0017] Based on the preprocessed friction torque data, the maximum transient friction torque deviation and engagement time deviation under the m-th running-in condition are obtained;

[0018] The overall friction torque and engagement time deviation under m running conditions in two cycles are evaluated. If the first preset condition is met, the running dynamic requirements are met.

[0019] Optionally, the maximum transient frictional torque deviation and engagement time deviation are respectively:

[0020]

[0021] Where ξ(m) and τ(m) represent the maximum transient friction torque deviation and engagement time deviation under the m-th working condition, and the subscript n-2 indicates the n-2th and nth cycles; M(t) represents the transient torque, and t1 and t0 represent the engagement end time and engagement start time, respectively.

[0022] The first preset condition is:

[0023]

[0024] Where m represents the break-in period.

[0025] Optionally, the assessment of micromorphological consistency includes:

[0026] Obtain the morphological feature parameters of the friction sample; wherein, the morphological feature parameters include: three-dimensional morphological roughness, surface profile arithmetic mean, surface skewness coefficient, and surface kurtosis coefficient.

[0027] If the morphological feature parameters meet the second preset conditions, the break-in period is considered to be over; otherwise, the friction sample is subjected to a cyclic test.

[0028] Optionally, the second preset condition is:

[0029]

[0030] Among them, S q For three-dimensional topography roughness, S a For the arithmetic mean of the surface profile, S sk S is the surface skewness coefficient.ku is the surface kurtosis coefficient.

[0031] Optionally, estimating the coefficient of friction and predicting the remaining number of runs during the break-in process of the second friction element includes:

[0032] Obtain the average coefficient of friction under preset operating conditions after the break-in period of friction elements of the same size. The preset working condition is a cyclic working condition that matches the clutch speed and oil pressure characteristics during the normal gear shifting process of a car.

[0033] The second friction element is subjected to the aforementioned cyclic test to obtain the average coefficient of friction under the preset working conditions during each cycle of the second friction element's break-in process. j represents the number of iterations;

[0034] A break-in period prediction model was constructed using the Long Short-Term Memory (LSTM) network method.

[0035] Using the average coefficient of friction Pre-train the break-in condition prediction model;

[0036] Using the average coefficient of friction Validate the break-in period prediction model;

[0037] Using the validated break-in state prediction model and preset break-in termination conditions, the evolution of the friction coefficient and the prediction of the remaining break-in cycles are estimated.

[0038] Optionally, the average friction coefficient is obtained. for:

[0039] Extract the instantaneous friction torque data of the preset working condition, and calculate the instantaneous friction coefficient of the preset working condition in the nth cycle;

[0040] The average friction coefficient is calculated based on the instantaneous friction coefficient.

[0041] Optionally, the instantaneous coefficient of friction is:

[0042]

[0043] Among them, M 2(t,n) Let s be the instantaneous friction torque under the second working condition in the nth cycle, s be the number of friction pairs, μ2(t,n) represent the instantaneous friction coefficient, p represent the contact pressure, and r be the instantaneous friction torque under the second working condition. o r i These represent the outer diameter and inner diameter of the friction element, respectively.

[0044] The average coefficient of friction for:

[0045]

[0046] Where t2 is the time (t0~t1) required for the clutch to complete one engagement under operating condition 2.

[0047] Optionally, the preset break-in termination condition is:

[0048]

[0049] in, This represents the predicted average friction coefficient. J represents the average friction coefficient at the end of the break-in period, and j′ represents the total number of break-in cycles required.

[0050] The beneficial effects of this invention are as follows:

[0051] This invention compares the overlap of instantaneous friction torques under various operating conditions in the (n-2)th and nth cycles, as well as the degree of deviation in engagement time, to evaluate the consistency of the running-in dynamics of the friction elements. Then, it compares the stability of the actual morphological characteristics of the friction elements to determine the number of cycles in the running-in stage and the end time of the running-in.

[0052] The method of this invention can clearly define the break-in end time, improve the break-in efficiency and quality of friction plates, and reduce the labor cost of the break-in process. For friction elements of the same material and size series, there are definite and accurate break-in conditions and number of cycles. For friction elements of the same material but different size series, a machine learning-based method for predicting the number of break-in cycles is proposed. This method does not require judgment of break-in dynamic consistency and actual morphological stability; it directly predicts the required number of break-in cycles based on the average friction test data from previous break-in tests. Attached Figure Description

[0053] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 This is a schematic diagram of an efficient running-in method for friction elements according to an embodiment of the present invention;

[0055] Figure 2 This is a schematic diagram of transient friction torque under a certain working condition according to an embodiment of the present invention;

[0056] Figure 3 This is a schematic diagram of the average friction coefficient under each cycle condition in working condition 2 of this embodiment of the invention;

[0057] Figure 4 The friction torque diagrams are shown for the 87th and 89th cycles of this embodiment of the invention.

[0058] Figure 5 The image shows the morphology of the friction plate according to an embodiment of the present invention; where (a) is before break-in and (b) is after break-in. Detailed Implementation

[0059] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0060] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0061] like Figure 1 As shown, this embodiment proposes a high-efficiency break-in method for friction elements, including:

[0062] The preset running-in conditions are imported into the pre-built clutch test control system, and the friction sample is subjected to cyclic test to extract the friction torque data of the friction sample under the running-in conditions in the preset number of cycles.

[0063] Specifically, in this embodiment, a test specimen, i.e., a friction specimen, is first prepared: based on the dimensional characteristics of the clutch under test, the inner and outer diameters are r and r, respectively. i r o To prepare the mating steel sheets and friction plates.

[0064] Then, the test bench, i.e., the clutch test control system, is constructed: The drive motor, speed and torque sensor 1, clutch housing, speed and torque sensor 2, and moment of inertia are connected sequentially. The clutch contains two mating steel plates and one friction plate, with a friction pair number N of 2. Speed ​​and torque sensors 1 and 2 measure the speed and torque of the clutch's active and passive ends, respectively. The hydraulic system controls the clutch piston oil pressure p and the lubrication between the friction elements. The moment of inertia is 2 kg·m. 2 .

[0065] The motor speed is n1(t), where t represents time; similarly, the torques at the driven end are n2(t) and M(t).

[0066] Next, several break-in conditions were set. In order to simulate the clutch break-in performance under full oil pressure and full speed difference range, the break-in cycle conditions are set as shown in Table 1:

[0067] Table 1

[0068] Operating conditions 1 2 3 4 5 6 7 8 p(MPa) 0.6 1.2 1.8 2.4 1.2 1.2 1.2 1.2 <![CDATA[n1(r / min)]]> 1200 1200 1200 1200 600 1200 1800 2400

[0069] The above break-in cycle conditions were imported into the clutch test control system, controlling the motor speed and the piston oil pressure characteristics of the hydraulic system. Each condition lasted 30 seconds and included engagement, disengagement, and cooling processes. Additionally, the number of cycles n for each condition was recorded, and the sampling frequency Δt for friction torque was set to 0.05 seconds.

[0070] The friction torque data is verified for consistency. If the consistency is satisfied, the microstructure consistency is evaluated. If the consistency is not satisfied, the friction sample is subjected to cyclic testing.

[0071] Furthermore, the friction torque data consistency verification includes:

[0072] Extract the friction torque data of the friction sample under m running conditions in the (n-2)th and nth cycles;

[0073] The friction torque data is preprocessed, including deburring and filtering.

[0074] Based on the preprocessed friction torque data, the maximum transient friction torque deviation and engagement time deviation under the m-th running-in condition are obtained;

[0075] The overall friction torque and engagement time deviation under m running conditions in two cycles are evaluated. If the first preset condition is met, the running dynamic requirements are met.

[0076] Furthermore, the assessment of micromorphological consistency includes:

[0077] Obtain the morphological feature parameters of the friction sample; among which, the morphological feature parameters include: three-dimensional morphological roughness, surface profile arithmetic mean, surface skewness coefficient, and surface kurtosis coefficient.

[0078] If the morphological characteristic parameters meet the second preset formula, the break-in period is considered complete; otherwise, the friction sample is subjected to cyclic testing.

[0079] Specifically, in this embodiment, firstly, a cyclic condition 2 that matches the clutch speed and oil pressure characteristics during a normal gear shift in a car is selected. The instantaneous friction torque data of condition 2 is extracted, and the instantaneous friction coefficient of condition 2 in the nth cycle is calculated:

[0080]

[0081] In the formula, M2(t,n) is the instantaneous friction torque under condition 2 in the nth cycle, and s is the number of friction pairs, s=2.

[0082] The average friction coefficient can be calculated from the instantaneous friction coefficient and is defined as follows:

[0083]

[0084] In the formula, t2 is the time (t0~t1) required for the clutch to complete one engagement under operating condition 2. A schematic diagram of transient friction torque is shown below. Figure 2 As shown.

[0085] Then, the friction torque data of the clutch under the above 8 operating conditions in the (n-2)th and nth cycles are extracted, and after deburring and filtering, the consistency of friction torque is verified: the consistency of transient friction torque under the above 8 operating conditions in the (n-2)th and nth cycles is compared. The specific method is as follows:

[0086] ① Compare the degree of deviation of transient friction torque in each working condition under two cycles, and determine the initial growth point time t0 of the two friction torque curves under the same working condition in two cycles. n-2 ,t0 n Then, the deviation of the transient friction torque at each subsequent sampling frequency Δt is calculated until the engagement termination time min(t1) is reached. n-2 ,t1 n The subscript n-2, where n represents the (n-2)th and nth iterations. Let the comparison nodes under the two iterations be: the (n-2)th iteration t... n-2 =t0 n-2 +Δt,t0 n-2 +2Δt,t0 n-2 +3Δt,……,min(t1 n-2 ,t1 n ); the (n-2)th iteration t n =t0 n +Δt,t0 n +2Δt,t0 n +3Δt,……,min(t1 n-2 ,t1 n );

[0087] The maximum transient frictional torque deviation and engagement time deviation under the m-th operating condition are obtained.

[0088]

[0089] ② Evaluate the overall friction torque and engagement time deviation under 8 operating conditions in two cycles. If the following formula is met, the break-in dynamic requirements are met; if the following formula is not met, the cycle test needs to be continued.

[0090]

[0091] Next, a microstructure consistency assessment is performed. If the friction torque consistency requirement is met, the friction plate is removed, and the three-dimensional surface roughness of the friction plate is obtained using an optical profilometer to determine the scanning area. The three-dimensional surface roughness S of the friction plate is obtained by performing multiple scans at different positions. q Surface profile arithmetic mean S a Surface skewness coefficient S sk and surface peak state coefficient S ku If the morphological characteristics satisfy the following expert empirical formula, the break-in period can be considered complete. If not, the cycle experiment in step three should be continued.

[0092]

[0093] After evaluating the consistency of the microstructure, the number of cycles n is recorded. Based on the number of cycles n, friction elements of the same size series are run-in, and the average coefficient of friction is extracted.

[0094] Specifically, in this embodiment, if the consistency of the microstructure is evaluated, the number of cycles n is recorded and input into the control system. Then, the running-in test of the friction elements of the same size series can be completed by directly using several cyclic conditions to perform n cycles.

[0095] The friction element is run-in based on the final number of cycles, i.e., the number of cycles n. For the first friction element of the same friction material and size series, only the final number of cycles is required. For the second friction element of the same friction material but different size series, the second friction element is run-in based on the final number of cycles, the average friction coefficient is extracted, and the friction coefficient during the second friction element's run-in process and the remaining number of run-in cycles are predicted based on the average friction coefficient.

[0096] Furthermore, the estimation of the friction coefficient and the prediction of the remaining number of running-in cycles for friction elements of the same friction material but different size series include:

[0097] Obtain the average coefficient of friction under preset operating conditions after the break-in period of friction elements of the same size. The preset working condition is a cyclic working condition that matches the clutch speed and oil pressure characteristics during the normal gear shifting process of a car.

[0098] Cyclic tests were conducted on friction elements of the same friction material but different size series to obtain the average coefficient of friction under preset working conditions during the break-in process of friction elements of different size series in each cycle. j represents the number of iterations;

[0099] A break-in period prediction model was constructed using the Long Short-Term Memory (LSTM) network method.

[0100] Using the average coefficient of friction Pre-train the break-in condition prediction model;

[0101] Using the average coefficient of friction Validate the break-in period prediction model;

[0102] Using the validated break-in state prediction model and preset break-in termination conditions, the evolution of the friction coefficient and the prediction of the remaining break-in cycles are estimated.

[0103] Specifically, in this embodiment, the average friction coefficient data of the clutch under normal operating conditions of 1.2 MPa and 1200 r / min are extracted. A machine learning model was built to estimate the friction coefficient and predict the remaining number of runs for friction elements of the same friction material but different sizes during the break-in process.

[0104] ① Collect the average friction coefficient data of the clutch after the break-in period under the above-mentioned size conditions. Next, outlier removal and missing value handling are performed to ensure data quality.

[0105] ② Repeat steps 1-4 above to conduct a break-in test on the new size series clutches, and obtain the average friction coefficient under operating condition 2 in each cycle during the break-in process of the new size series clutches. j represents the number of iterations.

[0106] ③ A break-in period prediction model is constructed using the Long Short-Term Memory (LSTM) network method. The LSTM prediction model mainly consists of an input layer, a hidden layer, and an output layer. The input layer is the previously obtained... The hidden layer consists of multiple LSTM units (the number of layers and units can be adjusted according to data complexity); the output layer represents the remaining number of run-in cycles for the friction element at this size. Offline data is used. Pre-training the LSTM model allows it to initially acquire the historical data evolution characteristics of the break-in process.

[0107] ④ Using partial online friction coefficient data from a friction element of a certain size that is currently in use. As a validation set, the model is adaptively optimized based on new data to avoid overfitting. Finally, the optimized model and preset break-in termination conditions are applied to estimate the evolution of the friction coefficient in future break-in cycles and predict the remaining break-in cycles, and the model performance is evaluated and optimized.

[0108] Termination condition definition: Based on the predicted average friction coefficient and Average coefficient of friction at the end of the break-in period By comparing the values ​​and assessing the degree of deviation, we determine whether the break-in period for this size series of clutches should be terminated. The termination conditions are as follows:

[0109]

[0110] If the above conditions are met, the total number of break-in cycles required can be obtained as j′.

[0111] This embodiment applies machine learning methods to propose a method for predicting the number of running-in cycles for friction elements of different sizes based on the average friction coefficient. By using the average friction experimental data from previous running-in tests, the required number of running-in cycles can be directly predicted.

[0112] The following describes the implementation steps of a high-efficiency break-in method for friction elements in this embodiment, using specific data:

[0113] 1. Preparation of test specimens: Based on the dimensional characteristics of the clutch under test, mating steel plates and friction plates were prepared. The mating steel plates were 2mm thick, with a surface roughness of R08 and a hardness of HRC38-42. The friction plate core plate was 1.5mm thick, and the friction layer was 0.5mm thick. The inner and outer diameters r of the mating steel plates and friction plates were... i r o The measurements are 0.06m and 0.073m respectively.

[0114] 2. Set up the test bench.

[0115] 3. Set up the break-in conditions and conduct tests.

[0116] 4. Record the average friction coefficient under each cycle condition of operating condition 2, such as... Figure 3 As shown.

[0117] 5. Friction Torque Consistency Verification: By comparing the transient friction torque consistency under the above 8 conditions in the 87th and 89th cycles, 16 friction torque curves were obtained under the two cycles, as shown below. Figure 4 As shown in Table 2, the maximum transient frictional torque deviation and engagement time deviation under each working condition are solved, which meet the running-in dynamic requirements.

[0118] Table 2 Friction Torque Consistency Evaluation Table

[0119]

[0120]

[0121] 6. Remove the friction pad and use an optical profilometer to obtain the three-dimensional surface roughness of the friction pad, determining the scanning area to be 1.750mm × 2.333mm. Three locations with the same area are selected for scanning, and the resulting morphology parameters are shown in Table 3 below. If the morphology parameters meet the requirements for running-in morphology, the running-in is considered complete. The friction pad morphology image is shown below. Figure 5 As shown, where Figure 5 (a) is before break-in. Figure 5 (b) is after the break-in period.

[0122] Table 3 Morphological parameters

[0123]

[0124] 7. If the consistency of the microstructure is evaluated, record the number of cycles as 89 and input it into the control system. Then, the friction plate will be broken in by performing 89 cycles according to the several cyclic conditions shown in step 3.

[0125] This embodiment abandons the commonly used fixed-condition break-in test settings. Based on the characteristics of the operating conditions encountered during vehicle operation, such as crawling, acceleration, constant speed, high speed, and deceleration, the actual speed and oil pressure characteristics of the clutch are set. The break-in test is carried out through clutch engagement test, which can eliminate the influence of operating parameters on the break-in effect.

[0126] This embodiment compares the overlap of instantaneous friction torque under various operating conditions in the (n-2)th and nth cycles, as well as the degree of deviation in engagement time, to evaluate the consistency of the running-in dynamics of the friction element. Then, it compares the stability of the actual morphological characteristics of the friction element to determine the number of cycles in the running-in stage and the end time of the running-in.

[0127] This efficient break-in method clarifies the break-in end time, improves the break-in efficiency and quality of friction plates, and reduces labor costs associated with the break-in process. For friction elements of the same material and size series, there are definite and accurate break-in conditions and number of cycles. For friction elements of the same material but different size series, a machine learning-based method for predicting the number of break-in cycles is proposed. This method eliminates the need for consistency in break-in dynamics and determination of actual morphological stability, directly predicting the required number of break-in cycles based on average friction test data from previous break-in experiments.

[0128] The above embodiments are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Various modifications and improvements made by those skilled in the art to the technical solutions of the present invention without departing from the spirit of the present invention should fall within the protection scope defined by the claims of the present invention.

Claims

1. A method of efficient running-in of a friction element, characterized in that The method comprises the following steps: introducing a preset running-in condition into a pre-built clutch test control system, performing a cycle test on a friction sample, and extracting friction torque data of the friction sample under the running-in condition in a preset number of cycles; verifying the friction torque data for friction torque consistency, if the verification is satisfied, performing microscopic morphology consistency evaluation, if the verification is not satisfied, continuing to perform a cycle test on the friction sample; after the microscopic morphology consistency evaluation, recording a final cycle number; based on the final cycle number, performing running-in on the friction element; wherein for a first friction element of the same friction material and the same size series, only the final cycle number is needed for running-in; for a second friction element of the same friction material but different size series, based on the final cycle number, performing running-in on the second friction element, extracting an average friction coefficient, and based on the average friction coefficient, estimating the friction coefficient and predicting the remaining running-in number in the running-in process of the second friction element; the estimating of the friction coefficient and the predicting of the remaining running-in number in the running-in process of the second friction element comprise: Obtain the average friction coefficient of the same size friction element after the preset working condition of the end of running-in ; wherein the preset working condition is the cycle working condition in which the clutch speed and the oil pressure characteristics of the automobile regular gear shifting process are consistent. performing the cycle test on the second friction element to obtain the average friction coefficient of the preset working condition under each cycle in the running-in process of the second friction element j represents the number of cycles; building a running-in state prediction model by using a long short-term memory network method; Using average friction coefficient The running-in state prediction model is pre-trained; Using average friction coefficient The running-in state prediction model is verified; using the verified running-in state prediction model and a preset running-in termination condition to estimate the evolution of the friction coefficient and predict the remaining running-in number in the future running-in cycle.

2. The friction element high efficiency running-in method according to claim 1, characterized by, The friction sample comprises a pair of steel sheets and a friction plate. The cycle test on the friction sample comprises: performing n cycles on the friction sample, each cycle comprising m running-in conditions; wherein the running-in conditions comprise: a first condition of 0.6 MPa and 1200 r / min, a second condition of 1.2 MPa and 1200 r / min, a third condition of 1.8 MPa and 1200 r / min, a fourth condition of 2.4 MPa and 1200 r / min, a fifth condition of 1.2 MPa and 600 r / min, a sixth condition of 1.2 MPa and 1200 r / min, a seventh condition of 1.2 MPa and 1800 r / min, and an eighth condition of 1.2 MPa and 2400 r / min.

3. The high efficiency running-in method of a friction element according to claim 2, characterized in that, The friction torque data verification comprises: extracting friction torque data of the friction sample under m running-in conditions in the n-2th cycle and the n th cycle; preprocessing the friction torque data; wherein the preprocessing comprises deburring and filtering; based on the preprocessed friction torque data, obtaining a maximum transient friction torque deviation and a joint time deviation under the m th running-in condition; evaluating the overall friction torque and joint time deviation of m running-in conditions in two cycles, if a first preset condition is satisfied, the running-in dynamic requirement is met.

4. The method for efficient running-in of a friction element according to claim 3, characterized in that The maximum transient friction torque deviation and the joint time deviation are respectively: wherein, , Mm(t) represents the maximum transient friction torque deviation and engagement time deviation under the mth working condition, the subscripts n-2, n represent the n-2, n times of circulation; M(t) represents the transient torque, t1, t0 represent the time of the end of engagement and the initial engagement time, respectively. The first preset condition is: wherein m represents the running-in condition.

5. The method of claim 1, wherein The microscopic morphology consistency evaluation comprises: obtaining morphology characteristic parameters of the friction sample; wherein the morphology characteristic parameters comprise three-dimensional morphology roughness, surface profile arithmetic mean, surface skewness coefficient, and surface kurtosis coefficient: If the morphological characteristic parameter meets a second preset condition, it is determined that the running-in is ended, and if the second preset condition is not met, the cyclic test on the friction sample is continued.

6. The friction element high efficiency running-in method according to claim 5, characterized by, The second preset condition is: wherein S q is the three-dimensional profile roughness, S a is the arithmetical mean of the surface profile, S sk is the surface skewness coefficient, S ku is the surface kurtosis coefficient.

7. The high efficiency running-in method of a friction element according to claim 1, characterized by, obtaining the average friction coefficient is: Extracting the instantaneous friction torque data of the preset working condition, and calculating the instantaneous friction coefficient of the preset working condition under the n th cycle; calculating the average friction coefficient based on the instantaneous friction coefficient .

8. The high efficiency running-in method of a friction element according to claim 7, characterized in that, The instantaneous friction coefficient is: wherein M 2(t,n) is the instantaneous friction torque of the second operating condition in the n-th cycle, s is the number of friction pairs, denotes the instantaneous friction coefficient, denotes the contact pressure, , denote the outer diameter and the inner diameter of the friction element, respectively. The average friction coefficient is: wherein, is the time required for the clutch to complete one engagement under the second operating condition.

9. The high efficiency running-in method of a friction element according to claim 7, characterized by, The preset running-in termination condition is: wherein, represents the predicted average friction coefficient, represents the average friction coefficient at the end of the break-in, represents the total number of break-in required.

Images