Friction testing device and method based on artificial intelligence adaptive leveling

Abstract

The application provides a kind of friction test device and method based on artificial intelligence self-adapting leveling, it is related to friction and wear test technical field.The device includes support frame, first drive module, second drive module, stress monitoring module, attitude adjusting mechanism, signal acquisition module and control system.The stress monitoring module is used to obtain the multi-point stress distribution data of friction interface in real time.The attitude adjusting mechanism is used to carry the lower friction material and adjust its space attitude in real time.The control system includes artificial intelligence learning unit and control algorithm unit;Artificial intelligence learning unit is used to predict the interface stress variation trend based on historical stress data;Control algorithm unit is used to calculate attitude compensation instruction according to the prediction result and current real-time stress distribution data, and drive attitude adjusting mechanism to act.The method realizes the dynamic, accurate and self-adapting uniformization regulation and control of friction interface stress distribution through real-time monitoring, prediction and active adjustment.

Description

Technical Field

[0001] This invention relates to the field of friction and wear testing technology, and more specifically, to a friction testing device and method based on artificial intelligence adaptive leveling. Background Technology

[0002] Friction and wear testing is an important method for evaluating the service performance of materials. During the test, the uniformity of stress distribution at the friction interface directly affects the material contact state and wear behavior, and is a key factor in ensuring the accuracy and repeatability of test results. However, in actual tests, due to factors such as uneven microstructure of the sample surface, processing and assembly errors, initial contact deviations are common between friction pairs. Under normal load, this deviation leads to a non-uniform distribution of stress at the contact interface, causing local stress concentration and abnormal wear, resulting in test results that deviate from the true tribological properties of the material.

[0003] Friction and wear testing is an important method for evaluating the service performance of materials and is widely used in the development of high-energy-load friction pairs in rail transit braking systems, aerospace components, and other applications. In these tests, the uniformity of stress distribution at the friction interface directly affects the material contact state and wear behavior, and is a key factor in ensuring the accuracy and repeatability of test results. However, in actual testing, due to factors such as accumulated machining errors of components, assembly gaps, and uneven microstructure of the sample surface, initial contact deviations are common between friction pairs. Under normal loads, these deviations will lead to a non-uniform distribution of stress at the contact interface, resulting in localized stress concentration and abnormal wear, causing the test results to deviate from the true tribological properties of the material.

[0004] Existing friction testing devices typically employ rigid support structures, relying primarily on manual, empirical static leveling before the test to adjust the contact posture. This method lacks a real-time feedback mechanism based on interface stress distribution, making it difficult to guarantee leveling accuracy and repeatability. Furthermore, static leveling is only effective for the initial state and cannot adapt to dynamic changes in the contact state during the test. When the friction pair undergoes reciprocating or rotational motion, factors such as system vibration, thermal deformation caused by frictional heat, and uneven material wear can lead to continuous drift in the contact posture, causing the initial leveling state to quickly fail.

[0005] Under current technological conditions, real-time leveling for such dynamic evolution remains a blank slate. Existing friction testing equipment generally lacks an active attitude adjustment mechanism. Even simple control logic based on deviation feedback exists in some precision machining fields, but it is completely unable to cope with the complex time-varying characteristics in friction testing. The main reason is that existing adjustment methods generally lack a deep understanding of the evolution characteristics of stress distribution, and existing simple feedback mechanisms are subject to significant phase lag due to mechanical response delays, making it difficult to achieve instantaneous compensation.

[0006] Therefore, developing an intelligent friction testing equipment that can perceive the evolution of interface stress in real time and has the ability to predict phase advance, and realize adaptive leveling of contact state during friction, is of great significance for building a high-confidence experimental platform. Summary of the Invention

[0007] The purpose of this invention is to provide a friction testing device and method based on artificial intelligence adaptive leveling, which addresses the shortcomings of existing technologies and solves the problems mentioned in the background.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A friction testing device based on artificial intelligence adaptive leveling includes a support frame; a first drive module disposed on the side wall of the support frame; a second drive module disposed at the bottom of the support frame; a stress monitoring module installed at the power output end of the first drive module, used to clamp the friction material and apply a normal load, and to acquire multi-point stress distribution data of the friction interface in real time; an attitude adjustment mechanism disposed on the second drive module, used to support the friction material and adjust its spatial attitude in real time; a signal acquisition module electrically connected to the stress monitoring module, used to acquire stress distribution data; and a control system communicatively connected to the signal acquisition module, the first drive module, the second drive module, and the attitude adjustment mechanism, the control system including an artificial intelligence learning unit and a control algorithm unit; the artificial intelligence learning unit is used to predict the future stress change trend of the interface based on historical stress change data of the friction interface and output the prediction result; the control algorithm unit is connected to the artificial intelligence learning unit and the signal acquisition module respectively, used to calculate attitude compensation commands based on the prediction results and the real-time stress distribution data at the current moment, and drive the attitude adjustment mechanism to achieve dynamic leveling of the stress distribution of the friction interface.

[0009] Furthermore, the stress monitoring module includes a force sensor; an adapter extension rod, the upper end of which is connected to the force sensor; an upper friction material mounting base, the top of which is detachably connected to the lower end of the adapter extension rod for clamping the upper friction material, the bottom surface of which has a placement groove; and a stress monitoring device, which is located in the placement groove on the bottom surface of the upper friction material mounting base for contacting the upper friction material to monitor the stress distribution at the friction interface, the stress monitoring device including an array strain sensor and a protective layer covering the sensitive surface of the array strain sensor.

[0010] Furthermore, the upper friction material mounting base has symmetrical guide rail grooves on both sides that communicate with the placement groove. The grooves are equipped with centering and clamping devices to simultaneously center, clamp, and release the upper friction material.

[0011] Furthermore, the attitude adjustment mechanism includes a rotary drive unit for driving the lower friction material to rotate; at least three sets of linear adjustment units are arranged symmetrically around the rotary drive unit, and each set of linear adjustment units is connected to the rotary drive unit through a connecting rod, for independently adjusting the edge height of the rotary drive unit, thereby changing the spatial attitude of the lower friction material.

[0012] Furthermore, the linear adjustment unit includes a fixed base, a driving component, a synchronous belt drive pair, a guide slide, an inclined gain slide, and a driven adapter block. The driving component, synchronous belt drive pair, and guide slide are all mounted on the fixed base. The synchronous belt drive pair is driven by the driving component, and the slider of the guide slide is fixedly connected to the synchronous belt drive pair. The inclined gain slide is mounted on the slider of the guide slide, and its upper surface is inclined. The driven adapter block is slidably connected to the upper surface of the inclined gain slide block, and is fixedly connected to the lower end of the connecting rod.

[0013] Furthermore, the first drive module includes a Z-axis drive unit fixed to the support frame; a crossbeam with both ends rigidly connected to the power output ends of the two Z-axis drive units; and an X-axis drive unit and a Y-axis drive unit sequentially integrated on the lower end face of the crossbeam to form a cross-shaped orthogonal superimposed structure.

[0014] Furthermore, the artificial intelligence learning unit is used to acquire the stress distribution characteristic data of the friction interface transmitted by the signal acquisition module in real time, and to construct a stress distribution mapping model through a preset neural network. The artificial intelligence learning unit uses the mapping model to evaluate the interface stability under the current working condition, and outputs the predicted value of the interface stress distribution at the next moment in real time, as input data for the artificial intelligence advanced intervention item in the control algorithm unit.

[0015] Furthermore, the control algorithm unit employs a control algorithm combining model predictive control and iterative learning control to generate attitude compensation control quantities based on the current friction interface stress distribution data; it is also used to calculate the attitude compensation command for the next moment. The core control law is:

[0016] in, This is for the displacement compensation command generated in the next moment. To represent the lead time for response, For instantaneous feedback items, For proportional gain, This represents the real-time deviation between the measured stress center and geometric center of the friction interface at the current moment. For cross-period learning terms, The learning gain determines the weights by which the system extracts historical experience. Forgetting factor, for the former The data from each period is weighted, with periods closer to the current time having a higher weight. Corresponding to the previous Each cycle is in the same spatial coordinates Stress characteristics at the location, For advanced predictions by artificial intelligence, The function is the predicted value of the interface stress distribution at the next time step, output by the artificial intelligence learning unit based on the stress distribution mapping model. The predicted values ​​are converted into predicted displacement values ​​for proactive intervention, which are used to output predicted displacement commands before interface instability occurs.

[0017] Furthermore, the control system also includes a data processing module and a safety constraint mechanism, which are used to filter and smooth the stress distribution data collected by the stress monitoring module, and to constrain the attitude compensation control quantity output by the control algorithm unit.

[0018] This application also provides a friction testing method based on artificial intelligence adaptive leveling. The aforementioned friction testing device includes the following steps: The stress distribution data at multiple points on the friction interface is obtained through the stress monitoring module; By using the artificial intelligence learning unit in the control system, the trend of interface stress change is predicted based on stress distribution data. The attitude compensation command is calculated by the control algorithm unit in the control system based on the predicted trend and real-time stress distribution data. The drive attitude adjustment mechanism adjusts the spatial attitude of the lower friction material according to the attitude compensation command in order to achieve dynamic homogenization of the stress distribution at the friction interface.

[0019] The present invention has at least the following advantages or beneficial effects: 1. By integrating an array of strain sensors into a stress monitoring module, multi-point synchronous measurement and real-time signal feedback of the interface contact stress distribution are achieved, providing an objective and accurate data basis for leveling control, and also providing an unprecedented experimental observation method for the study of friction mechanisms.

[0020] 2. To address the inherent discrepancy between the instantaneous nature of friction interface evolution and the lag in mechanical adjustment, this invention utilizes a prediction algorithm built into the control system to achieve pre-intervention with phase advance. By employing the technique of "compensating for delay with prediction," the inherent physical response time of the mechanical transmission and actuator is effectively offset, ensuring that attitude cancellation can be completed instantly upon interface stress instability. This solves the problems of adjustment failure and system oscillation caused by phase lag in traditional feedback modes.

[0021] 3. According to research needs, the present invention can preset specific non-uniform stress distribution targets. The device can automatically calculate and drive the leveling mechanism to reach the target contact posture, thereby transforming the contact state from an interference variable that needs to be suppressed into an experimental parameter that can be precisely set, which greatly expands the experimental capabilities in cutting-edge fields such as simulating complex service conditions and studying the influence mechanism of stress distribution.

[0022] 4. The device has standardized mechanical and electrical interfaces. The first drive module can quickly mount various functional modules such as scratch, indentation, high temperature environment and in-situ observation. The second drive module can be adapted to different motion forms and fixtures, together forming an integrated high-end tribological testing platform with flexible configuration, which significantly improves the utilization rate and scientific research adaptability of the equipment. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A schematic diagram of the overall structure of a friction testing device and method based on artificial intelligence adaptive leveling provided by the present invention; Figure 2 This is a schematic diagram of the structure of the first drive module and the second drive module provided by the present invention; Figure 3 This is a schematic diagram of the structure of the clamping base provided by the present invention; Figure 4 This is a schematic diagram of the structure of the upper friction material mounting base provided by the present invention; Figure 5 This is also a schematic diagram of the structure of the upper friction material mounting base provided by the present invention; Figure 6 This is a schematic diagram of the stress monitoring device provided by the present invention; Figure 7 A schematic diagram of the attitude adjustment mechanism provided by the present invention; Figure 8 This is a schematic diagram of the structure of the linear adjustment module provided by the present invention; Figure 9 This is a schematic diagram of the control process of the artificial intelligence module and control algorithm unit used in this invention.

[0025] Icons: 1. Support frame; 2. First drive module; 21. Z-axis drive unit; 22. Crossbeam; 23. X-axis drive unit; 24. Y-axis drive unit; 25. Module adapter base; 3. Second drive module; 31. Horizontal linear guide drive unit; 4. Stress monitoring module; 41. Force sensor; 42. Adapter extension rod; 43. Upper friction material mounting base; 431. Connecting groove; 432. Fastener; 433. Placement groove; 434. Guide rail groove; 435. Centering clamping device; 436. Bidirectional wire 44. Lever mechanism; 441. Stress monitoring device; 442. Arrayed strain sensor; 443. Protective layer; 5. Attitude adjustment mechanism; 51. Rotary drive unit; 52. Linear adjustment unit; 521. Fixed base; 522. Drive component; 523. Synchronous belt drive pair; 524. Guide slide bar; 525. Inclined gain slide seat; 526. Driven adapter block; 53. Connecting rod; 54. Lower friction material fixture; 6. Information acquisition module; 7. Control system; 8. Upper friction material; 9. Lower friction material. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0027] Please refer to Figures 1 to 9As shown, a friction testing device based on artificial intelligence adaptive leveling includes a support frame 1, a first drive module 2, a second drive module 3, a stress monitoring module 4, an attitude adjustment mechanism 5, a signal acquisition module 6, and a control system 7. The first drive module 2 is located on the side wall of the support frame 1, and the second drive module 3 is located at the bottom of the support frame 1. The stress monitoring module 4 is installed at the power output end of the first drive module 2, used to clamp the friction material 8 and apply a normal load, and to acquire multi-point stress distribution data of the friction interface in real time. The attitude adjustment mechanism 5 is located on the second drive module 3, used to support the lowered friction material 9 and adjust its spatial attitude in real time. The signal acquisition module 6 is electrically connected to the stress monitoring module 4 and is used to acquire stress distribution data. The control system 7 is communicatively connected to the signal acquisition module 6, the first drive module 2, the second drive module 3, and the attitude adjustment mechanism 5, and includes an artificial intelligence learning unit 71 and a control algorithm unit 72. The artificial intelligence learning unit 71 is used to predict the future trend of interface stress change based on historical friction interface stress change data and output the prediction result. The control algorithm unit 72 is connected to the artificial intelligence learning unit 71 and the signal acquisition module 6, respectively. It calculates attitude compensation commands based on prediction results and real-time stress distribution data, and drives the attitude adjustment mechanism 5 to dynamically level the stress distribution at the friction interface. The device senses the stress distribution at the friction interface in real time through the stress monitoring module 4, and the signal acquisition module 6 transmits the data to the control system 7. The control algorithm unit 72, combining future trend predictions provided by the artificial intelligence learning unit 71, current real-time stress data, and historical experience across cycles, calculates advanced attitude compensation commands. These commands drive the attitude adjustment mechanism 5 to adjust the attitude of the lower friction material 9, thereby proactively compensating before stress unevenness occurs or intensifies. The closed-loop control process effectively overcomes the adjustment delay problem caused by mechanical response lag, achieving continuous homogenization of the stress distribution at the friction interface during the dynamic test process.

[0028] In practical use, the overall working process of the device is as follows: First, the operator clamps the upper friction material 8 and the lower friction material 9 onto the stress monitoring module 4 and the attitude adjustment mechanism 5, respectively. The control system 7 controls the first drive module 2 and the second drive module 3 to bring the two materials close to the critical contact position. Subsequently, based on the initial stress distribution data fed back by the stress monitoring module 4, the control system 7 drives the attitude adjustment mechanism 5 to perform static leveling, making the stress distribution tend to be uniform. After the formal test begins, the lower friction material 9 rotates under the drive of the attitude adjustment mechanism 5. At the same time, the stress monitoring module 4 continuously collects interface stress data, and the control system 7, combining artificial intelligence prediction and real-time feedback, continuously calculates and outputs attitude compensation commands, driving the attitude adjustment mechanism 5 to dynamically fine-tune the attitude of the lower friction material 9, thereby maintaining the stability and uniformity of the stress distribution at the friction interface throughout the entire test.

[0029] Please refer to Figure 2 As shown, the first drive module 2 specifically includes a Z-axis drive section 21, a crossbeam 22, an X-axis drive section 23, and a Y-axis drive section 24. The Z-axis drive section 21 is a linear guide rail, and two sections are provided, symmetrically arranged. The two Z-axis drive sections 21 are fixed to the support frame 1. Both ends of the crossbeam 22 are rigidly connected to the power output ends of the two Z-axis drive sections 21, respectively. The X-axis drive section 23 and the Y-axis drive section 24 are sequentially integrated onto the lower end face of the crossbeam 22, forming a cross-shaped orthogonal superposition structure. The X-axis drive section 23 is fixed to the crossbeam 22, and the Y-axis drive section 24 is mounted on the slider of the X-axis drive section 23. In this embodiment, both the X-axis drive section 23 and the Y-axis drive section 24 are linear guide rails. A module adapter base 25 is fixedly mounted on the slider of the Y-axis drive unit 24. The module adapter base 25 is configured with a standardized interface, which can quickly mount and adapt to various functional modules according to test requirements. The functional modules that can be quickly mounted on the module adapter base 25 include, but are not limited to, the following: a scratch test module, used to evaluate the bonding strength and scratch toughness of the coating or substrate on the material surface; a nano / micro indentation test module, used to measure the hardness, elastic modulus and other mechanical properties of the material at a small scale; a high-temperature environment module, which integrates a heating device and heat insulation design, and can perform friction, wear or indentation tests under high-temperature conditions; and an in-situ microscopic observation module, which integrates an optical microscope or confocal sensor, and can observe and record the wear morphology of the friction contact area in real time. Through the coordinated movement of the Z-axis drive unit 21, the X-axis drive unit 23 and the Y-axis drive unit 24, the first drive module 2 can achieve high-precision, multi-degree-of-freedom adjustment of the position of the upper friction material clamped on it in three-dimensional space.

[0030] The second drive module 3 is a set of horizontal linear guide drive units 31, which are used to drive the attitude adjustment mechanism 5 and the lower friction material 9 to perform high-precision reciprocating motion along a single axis.

[0031] Please refer to Figure 3 , Figure 4 , Figure 5 and Figure 6As shown, the stress monitoring module 4 specifically includes a force sensor 41, an adapter extension rod 42, an upper friction material mounting base 43, and a stress monitoring device 44. The upper friction material mounting base 43 is used to clamp the upper friction material 8, and its top is provided with a connecting groove 431 and a fastener 432. The upper end of the adapter extension rod 42 is connected to the force sensor 41, and its lower end is located within the connecting groove 431. It is detachably connected to the top of the upper friction material mounting base 43 via the fastener 432. A placement groove 433 is provided on the bottom surface of the upper friction material mounting base 43. The stress monitoring device 44 is disposed within the placement groove 433 on the bottom surface of the upper friction material mounting base 43, and is used to contact the upper friction material 8 to monitor the stress distribution at the friction interface. The stress monitoring device 44 includes an array strain sensor 441 and a protective layer 442 covering the sensitive surface of the array strain sensor 441. The protective layer 442 covers and is adhered to one side of the sensitive surface of the array strain sensor 441. The side of the array strain sensor 441 without the protective layer 442 is attached to the bottom of the placement groove 433. The protective layer 442 can contact the end face of the friction material 8, preventing direct contact between the array strain sensor 441 and the friction material 8 and causing wear. Furthermore, guide rail grooves 434, communicating with the placement groove 433, are symmetrically provided on both sides of the upper friction material mounting base 43. The guide rail grooves 434 have an isosceles trapezoidal cross-section. A centering clamping device 435 is provided within the guide rail groove 434 for synchronously centering, clamping, and releasing the upper friction material 8. The centering clamping device 435 is driven by a bidirectional lead screw mechanism 436. When the bidirectional lead screw mechanism 436 rotates, it drives the two centering clamping devices 435 to move synchronously towards or away from each other, thereby achieving automatic centering and clamping of the upper friction material 8 placed in the placement groove 433. This structure ensures that the installation position of the upper friction material 8 is precisely centered, reducing the initial stress unevenness introduced by clamping eccentricity. The bidirectional lead screw mechanism 436 can be driven to rotate by a manual handle or a micro motor to perform loading and unloading actions.

[0032] Please refer to Figure 7 and Figure 8As shown, the attitude adjustment mechanism 5 specifically includes a rotary drive unit 51 and at least three sets of linear adjustment units 52. The rotary drive unit 51 is a DD motor used to drive the lower friction material 9 to rotate. A lower friction material clamp 54 for holding the lower friction material 9 is provided on the top of the rotary drive unit 51. At least three sets of linear adjustment units 52 are arranged symmetrically around the outer periphery of the rotary drive unit 51. Each set of linear adjustment units 52 is connected to the rotary drive unit 51 through a connecting rod 53 and is used to independently adjust the edge height of the rotary drive unit 51, thereby changing the spatial attitude of the lower friction material 9. By independently controlling the extension and retraction of the three sets of linear adjustment units 52, the tilt angle of the rotary drive unit 51 and the lower friction material 9 it carries can be precisely adjusted to achieve multi-degree-of-freedom attitude adjustment. Specifically, the linear adjustment unit 52 includes a fixed base 521, a drive component 522, a synchronous belt drive pair 523, a guide slide rod 524, an inclined gain slide seat 525, and a driven adapter block 526. The drive component 522, the synchronous belt drive pair 523, and the guide slide 524 are all mounted on the fixed base 521. The synchronous belt drive pair 523 is driven by the drive component 522. The guide slide 524 includes a fixed seat and a slider. The fixed seat is fixedly mounted on the fixed base 521, and the slider is slidably connected to the fixed seat and fixedly connected to the synchronous belt drive pair 523. The inclined gain slide 525 is mounted on the slider of the guide slide 524, and the upper surface of the inclined gain slide 525 is inclined. The driven adapter block 526 is slidably connected to the upper surface of the inclined gain slide 525 through a groove, guide rail, or other means, and the driven adapter block 526 is fixedly connected to the lower end of the connecting rod 53. When the drive component 522 drives the synchronous belt transmission pair 523 to move the inclined plane gain sliding seat 525 horizontally via the slider of the guide slide rod 524, the driven adapter block 526 slides along the inclined plane, thereby accurately converting the horizontal displacement of the inclined plane gain sliding seat 525 into the vertical displacement of the driven adapter block 526, which is then transmitted to the rotary drive unit 51 through the connecting rod 53. The inclined plane gain structure can achieve high-precision displacement adjustment at the micron level and convert the horizontal movement of the drive component 522 into changes in vertical support force, with a compact structure and smooth transmission.

[0033] Please refer to Figure 9 As shown, to overcome the real-time deviation in adjustment caused by the lag in response and the pre-emptive motion state of the mechanical transmission system, this invention achieves closed-loop control based on the artificial intelligence learning unit and control algorithm unit integrated within the control system 7. The artificial intelligence learning unit acquires the friction interface stress distribution characteristic data transmitted by the information acquisition module in real time and constructs a stress distribution mapping model through a preset neural network. The artificial intelligence learning unit uses the mapping model to evaluate the interface stability under the current working condition and outputs the predicted interface stress distribution value for the next moment in real time. This serves as the input data for the AI-driven proactive intervention in the control algorithm unit. The control algorithm unit integrates the forward-looking capabilities of model predictive control with the periodic learning capabilities of iterative learning control, aiming to achieve proactive and adaptive compensation for the attitude of the friction interface.

[0034] The control algorithm unit deeply integrates the look-ahead capability of model predictive control with the periodic learning capability of iterative learning control. Based on the current stress distribution data of the friction interface and the predicted values ​​provided by the artificial intelligence learning unit, it pre-calculates the attitude compensation amount required to offset the stress shift at the next moment, generating the attitude compensation control quantity. Specifically, after the experiment starts, the system first enters a period of... During the sampling and learning phase of one reciprocating or rotating cycle, the control system 7 uses the information acquisition module 6 to sample at a fixed frequency. Continuously collect the following multi-dimensional time-series data: real-time spatial coordinates of the lower friction material. Normal load fed back by force sensor 41 and the two-dimensional stress field vector data fed back by the array strain sensor 441 Based on the aforementioned multi-dimensional time-series data, the control system 7 performs [measurements] at each sampling time. Calculate and output pose control commands for the next moment in real time. Its core control law formula is expressed as follows:

[0035] This refers to the displacement command generated by the system for the next moment, where This represents the response lead time. This parameter is preset based on the mechanical inertia of the drive component 522 and the synchronous belt drive pair 523. By reserving a compensation amount on the time axis, it ensures that the physical action of the attitude adjustment mechanism 5 can lead the positional requirements of the frictional motion, thus eliminating the drawback of the adjustment action being delayed. It serves as an instantaneous feedback term, forming the basis of closed-loop control, and is used to quickly correct real-time deviations and maintain system stability. For proportional gain, This represents the real-time deviation between the measured stress center and the geometric center of the specimen at the current moment. This is a cross-cycle learning term, used to learn and compensate for repeatability periodic errors caused by initial sample unevenness, fixture eccentricity, etc. The learning gain determines the weight of the system's extraction of historical experience; Forgetting factor (weighting coefficient), for the former The data from each period is weighted, with periods closer to the current time having higher weights. Corresponding to the previous Each cycle is in the same phase The stress characteristics at the location. This item aims to predict and eliminate repeatability periodic errors caused by initial unevenness of the sample, eccentricity of the fixture, etc. For advanced predictions by artificial intelligence, among which The function is the predicted value of the interface stress distribution at the next time step, output by the artificial intelligence learning unit based on the stress distribution mapping model. The predicted values ​​are converted into predicted displacement values ​​for proactive intervention, which are used to output predicted displacement commands before interface instability occurs.

[0036] During the implementation process, the control system 7 will synthesize control commands. The motion components of the three independent linear adjustment modules 52 in the attitude adjustment mechanism 5 are decoupled in real time. When the drive component 522 drives the inclined gain sliding seat 525 to move in advance according to the instruction, the vertical support height is finely adjusted at the instant before the friction pair moves to the high stress risk position through precise conversion of the inclined height.

[0037] In a further embodiment, the control system 7 also includes a data processing module and a safety constraint mechanism, used to filter and smooth the stress distribution data collected by the stress monitoring module 4, and to constrain the attitude compensation control quantity output by the control algorithm unit. The data processing module can filter out high-frequency noise in the signal and extract the true stress change trend. The safety constraint mechanism is used to ensure that the attitude compensation command does not exceed the mechanical travel limit of the attitude adjustment mechanism 5, or to trigger a protection program (such as stopping movement, lifting the load, etc.) when abnormal stress changes occur, to ensure the safety of the equipment and the sample.

[0038] Furthermore, the control system 7 can be equipped with multiple operating modes, such as static leveling mode, dynamic leveling mode, and autonomous programming control mode. In static leveling mode, the system mainly relies on instantaneous feedback and trend prediction for rapid leveling; in dynamic leveling mode, the system uses a complete composite control algorithm that includes cross-cycle learning terms; in autonomous programming mode, the operator is allowed to preset specific contact state target parameters, and the system drives the attitude adjustment mechanism 5 based on the composite control algorithm to form and maintain a specific stress distribution at the friction interface corresponding to the target parameters, thereby transforming the contact state into a precisely controllable experimental variable.

[0039] This application also provides a friction testing method based on artificial intelligence adaptive leveling, using the friction testing device described above, and the method includes the following steps: S1. Secure the upper friction material mounting base 43 to the end of the adapter extension rod 42 using fasteners 432. Then, place the upper friction material 8 in the centering clamping device 435 and drive the bidirectional lead screw mechanism 436 to bring it to a pre-clamped state that is initially centered and capable of slight movement. Place the lower friction material 8 in the lower friction material fixture 54 and lock it to the upper surface of the rotary drive unit 51 with bolts. Start the control system 7 and use the information acquisition module 6 to perform reference zeroing on the force sensor 41, the array strain sensor 441, and the displacement sensing units in the first and second drive modules to ensure that the multi-channel signal references are consistent.

[0040] S2. Manually or automatically control the first drive module 2 and the second drive module 3 to move the upper friction material 8 and the lower friction material 9 to the vicinity of the target contact area and then move them downwards to the critical position where they are about to make contact. Control the first drive module 2 to move downwards at a micro-feed speed until the normal resultant force fed back by the stress monitoring module 4 reaches the preset contact threshold, then stop feeding to complete the initial micro-touch loading.

[0041] S3. Control system 7 initiates static leveling mode to perform static and precise leveling of the initial contact position of the friction pair materials. In this mode, since the friction pair is in a stationary or quasi-stationary state, the system does not need to utilize data from historical motion cycles; therefore, the learning gain in the cross-cycle learning term is... L Set to 0. The algorithm only activates the instantaneous feedback term and the trend prediction term, and its control instructions are simplified to:

[0042] Among them, proportional gain Set to a higher value for rapid deviation correction; trend prediction function The system maintains its early warning capability for sudden changes in contact state. It calculates in real-time the height compensation required for the three support points to homogenize the stress distribution and converts this into corresponding horizontal displacement commands based on the motion relationship of the inclined plane gain mechanism. The control system 7 then drives the linear adjustment unit 52 until the real-time monitored stress distribution uniformity reaches a preset threshold. After the operator confirms the leveling effect is satisfactory, the system finally drives the centering clamping device 435 to completely lock the upper friction material 8, completing the precise calibration of the initial posture.

[0043] S4. After leveling, set the test parameters, including normal load and test duration. The motion mode can be selected as reciprocating or rotary. If reciprocating is selected, the motion stroke, frequency, and acceleration / deceleration time of the second drive module 3 need to be set; if rotary is selected, the rotation drive unit 51 needs to be rotated by command through the control system 7, and the angular velocity and acceleration / deceleration time need to be set.

[0044] S5. During the formal test, the control system 7 automatically switches to dynamic leveling mode to address the real-time challenges posed by relative motion and system inertia during the test. In this mode, the control system collects stress data in real time, calculates advance compensation commands through the artificial intelligence learning unit and control algorithm unit, and drives the attitude adjustment mechanism to dynamically adjust the attitude of the friction material to maintain the stability and uniformity of the stress distribution at the friction interface during the test.

[0045] S6. When the test ends or the protection condition is triggered, stop the movement and save the data.

[0046] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A friction testing device based on artificial intelligence adaptive leveling, characterized in that, include: Support frame; The first drive module is located on the side wall of the support frame; The second drive module is located at the bottom of the support frame; The stress monitoring module is installed at the power output end of the first drive module. It is used to clamp the friction material and apply a normal load, and to acquire multi-point stress distribution data of the friction interface in real time. The attitude adjustment mechanism is mounted on the second drive module and is used to support the lower friction material and adjust its spatial attitude in real time. The signal acquisition module is electrically connected to the stress monitoring module and is used to acquire stress distribution data; The control system is communicatively connected to the signal acquisition module, the first drive module, the second drive module, and the attitude adjustment mechanism. The control system includes an artificial intelligence learning unit and a control algorithm unit. The artificial intelligence learning unit is used to predict the trend of interface stress change in the future based on historical friction interface stress change data and output the prediction results. The control algorithm unit is connected to the artificial intelligence learning unit and the signal acquisition module respectively. It is used to calculate the attitude compensation command based on the prediction results and the real-time stress distribution data at the current moment, and drive the attitude adjustment mechanism to achieve dynamic leveling of the stress distribution at the friction interface.

2. The friction testing device based on artificial intelligence adaptive leveling according to claim 1, characterized in that, The stress monitoring module includes: Force sensor; An adapter extension rod is provided, the upper end of which connects to a force sensor. The upper friction material mounting base has its top detachably connected to the lower end of the adapter extension rod for clamping the upper friction material. The bottom surface of the upper friction material mounting base has a placement groove. A stress monitoring device is installed in a placement groove on the bottom surface of the upper friction material mounting base. It is used to contact the upper friction material to monitor the stress distribution at the friction interface. The stress monitoring device includes an array of strain sensors and a protective layer covering the sensitive surface of the array of strain sensors.

3. The friction testing device based on artificial intelligence adaptive leveling according to claim 2, characterized in that, The upper friction material mounting base has symmetrical guide rail grooves on both sides that communicate with the placement groove. The grooves are equipped with centering and clamping devices to synchronously center and clamp and release the upper friction material.

4. The friction testing device based on artificial intelligence adaptive leveling according to claim 1, characterized in that, The attitude adjustment mechanism includes: A rotary drive unit is used to drive the lower friction material to rotate. At least three sets of linear adjustment units, each equipped with a connecting rod, are arranged symmetrically around the outer periphery of the rotary drive unit. Each set of linear adjustment units is fixedly connected to the rotary drive unit through a corresponding connecting rod, and is used to independently adjust the edge height of the rotary drive unit, thereby changing the spatial attitude of the lower friction material.

5. The friction testing device based on artificial intelligence adaptive leveling according to claim 4, characterized in that, The linear adjustment unit includes a fixed base, a drive component, a synchronous belt drive pair, a guide slide, an inclined gain slide, and a driven adapter block. The drive component, the synchronous belt drive pair, and the guide slide are all mounted on the fixed base. The guide slide includes a fixed seat and a slider slidably connected to the fixed seat. The slider is fixedly connected to the synchronous belt drive pair. The inclined gain slide is mounted on the slider of the guide slide, and the upper surface of the inclined gain slide is inclined. The driven adapter block is slidably connected to the upper surface of the inclined gain slide, and the driven adapter block is fixedly connected to the lower end of the connecting rod.

6. The friction testing device based on artificial intelligence adaptive leveling according to claim 1, characterized in that, The first drive module includes: The Z-axis drive unit is fixed on the support frame; The crossbeam is rigidly connected at both ends to the power output ends of the two Z-axis drive units; The X-axis drive unit and the Y-axis drive unit are sequentially integrated on the lower end face of the crossbeam to form a cross-shaped orthogonal superposition structure.

7. The friction testing device based on artificial intelligence adaptive leveling according to claim 1, characterized in that, The artificial intelligence learning unit acquires the stress distribution characteristics data of the friction interface transmitted by the information acquisition module in real time, and constructs a stress distribution mapping model through a pre-set neural network. The artificial intelligence learning unit uses the mapping model to evaluate the interface stability under the current working condition and outputs the predicted value of the interface stress distribution at the next moment in real time. , as input data for the AI ​​advanced intervention items in the control algorithm unit.

8. The friction testing device based on artificial intelligence adaptive leveling according to claim 1, characterized in that, The control algorithm unit adopts a control algorithm that combines model predictive control and iterative learning control to generate attitude compensation control quantities based on the current friction interface stress distribution data. It is used to calculate the attitude compensation command for the next moment. The core control law is: in, This is for the displacement compensation command generated in the next moment. To represent the lead time for response, For instantaneous feedback items, For proportional gain, This represents the real-time deviation between the measured stress center and geometric center of the friction interface at the current moment. For cross-period learning terms, The learning gain determines the weights by which the system extracts historical experience. Forgetting factor, for the former The data from each period is weighted, with periods closer to the current time having a higher weight. Corresponding to the previous Each cycle is in the same spatial coordinates Stress characteristics at the location, For advanced predictions by artificial intelligence, The function is the predicted value of the interface stress distribution at the next time step, output by the artificial intelligence learning unit based on the stress distribution mapping model. The predicted values ​​are converted into predicted displacement values ​​for proactive intervention, which are used to output predicted displacement commands before interface instability occurs.

9. The friction testing device based on artificial intelligence adaptive leveling according to claim 1, characterized in that, The control system also includes a data processing module and a safety constraint mechanism, which are used to filter and smooth the stress distribution data collected by the stress monitoring module, and to constrain the attitude compensation control quantity output by the control algorithm unit.

10. A friction testing method based on artificial intelligence adaptive leveling, characterized in that, The friction testing apparatus as described in any one of claims 1 to 9 comprises the following steps: The stress distribution data at multiple points on the friction interface is obtained through the stress monitoring module; By using the artificial intelligence learning unit in the control system, the trend of interface stress change is predicted based on stress distribution data. The attitude compensation command is calculated by the control algorithm unit in the control system based on the predicted trend and real-time stress distribution data. The drive attitude adjustment mechanism adjusts the spatial attitude of the lower friction material according to the attitude compensation command in order to achieve dynamic homogenization of the stress distribution at the friction interface.

Images