A caster fatigue life test method and related system

Abstract

The application relates to the technical field of caster life, in particular to a caster fatigue life test method and a related system, which comprises the following steps: obtaining an initial state reference value of a to-be-tested caster, controlling the to-be-tested caster to continuously roll and obtaining multi-parameter continuous data corresponding to the initial state reference value; performing stage-by-stage identification on the multi-parameter continuous data, setting a specific feature point when a preset inflection point is identified, and recording the current rolling circle number, wherein the preset inflection point is a data point in the multi-parameter continuous data, at which a data change rate exceeds a preset threshold; performing comprehensive analysis on the multi-parameter continuous data after the specific feature point, and determining that the to-be-tested caster reaches a fatigue state and recording the total rolling circle number at this time when a preset condition is met. The application solves the defects of strong subjectivity and poor consistency caused by single macroscopic failure characteristics such as naked eye visibility in the prior art, and makes the determination of the fatigue life end value highly objective, quantifiable and traceable.

Description

Technical Field

[0001] This invention relates to the field of caster life testing technology, and more specifically, to a method and related system for testing the fatigue life of casters. Background Technology

[0002] Casters are mobile load-bearing components used in furniture, medical equipment, logistics equipment, and other applications. Their fatigue life is crucial to the overall safety and durability of these products. Currently, caster fatigue life testing primarily refers to standards such as ISO 22878. On a dedicated test bench, a constant or alternating radial load is applied to the caster, causing it to roll continuously along a fixed track or roller until specified failure characteristics appear. The total number of rolling revolutions at failure is then used as the fatigue life value. This method is widely used in the industry and provides a fundamental means for evaluating caster performance.

[0003] However, existing caster fatigue life testing methods generally suffer from overly simplistic fatigue failure criteria and focus on macroscopic endpoint failure characteristics, leading to a lack of objectivity and consistency in determining the final fatigue life value. Current standards or testing specifications typically use visually observable or extreme functional loss phenomena such as significant wheel deformation, bracket breakage, bearing jamming resulting in inability to roll, or a radial load reduction exceeding a certain fixed percentage as the sole termination condition. This method leads to significant differences in subjective interpretation of significant deformation or jamming among different testers, potentially resulting in several times the difference in fatigue life data for the same batch of casters in different laboratories. Furthermore, caster fatigue failure is a continuous and gradual process from micro-damage accumulation to rapid instability. Simply waiting for complete macroscopic functional loss to terminate the test often causes the recorded fatigue life value to deviate significantly from the actual safe and usable limit, resulting in an overestimation of product durability and posing serious safety hazards.

[0004] Therefore, there is an urgent need for a new method for testing the fatigue life of casters that can eliminate subjective judgment and objectively identify the stage based on the damage evolution process itself, so as to achieve the quantification and traceable determination of the final fatigue life value. Summary of the Invention

[0005] The main objective of this invention is to provide a method and system for testing the fatigue life of casters, aiming to overcome the technical problems of lack of objectivity and consistency in the determination of the final value of fatigue failure life in the prior art.

[0006] To address the aforementioned problems, this invention proposes a method for testing the fatigue life of casters, the method comprising: Obtain the initial state reference value of the caster under test, control the caster under test to roll continuously and obtain multi-parameter continuous data corresponding to the initial state reference value; The multi-parameter continuous data is identified in stages and levels. When a preset inflection point is identified, a specific feature point is set and the current number of scrolling cycles is recorded. The preset inflection point is a data point in the multi-parameter continuous data whose data change rate exceeds a preset threshold. A comprehensive analysis is performed on the continuous multi-parameter data following the specific feature point. When a preset condition is detected, the caster under test is determined to have reached a fatigue state, and the total number of rolling revolutions at this time is recorded. The preset condition is that the value of a specific parameter in the continuous multi-parameter data exceeds a preset fatigue life determination threshold.

[0007] Furthermore, the step of obtaining the initial state reference value of the caster under test includes: The initial state reference value is obtained by acquiring the force signal, radial runout and yaw amplitude and peak amplitude of the caster under test based on the force sensor, displacement sensor and acoustic emission sensor on the caster under test.

[0008] Furthermore, the step of controlling the caster under test to continuously roll and acquiring multi-parameter continuous data corresponding to the initial state reference value includes: The caster under test is controlled to roll continuously, and the number of rolls is obtained from the photoelectric sensor inside the caster under test. Obtain continuous data corresponding to the number of rolling revolutions of the initial state reference value, and divide the continuous data into data segments according to the preset number of revolutions; Based on each data segment, continuous load fluctuation data is calculated based on the triaxial force signal, continuous radial runout data is calculated based on the radial runout and yaw amplitude, and continuous acoustic emission cumulative energy data is calculated based on the yaw amplitude and peak amplitude, thus obtaining multi-parameter continuous data.

[0009] Furthermore, the step of performing staged, step-by-step identification of the multi-parameter continuous data includes: A sliding window with a preset number of bits is constructed to smooth the continuous radial runout data, resulting in a smoothed radial runout data curve. The local slope of the radial runout data curve is calculated according to the first preset number of cycles interval, and the average slope of multiple first preset number of cycles intervals is set as the first reference slope. When the local slopes of multiple consecutive adjacent intervals of the first preset number of cycles all exceed a preset multiple of the first reference slope and show a monotonically increasing trend, the starting number of rolling cycles of the first interval that meets the condition is set as the first feature point, and the number of rolling cycles corresponding to the first feature point is recorded.

[0010] Furthermore, the step of performing phased and progressive identification of the multi-parameter continuous data, setting a specific feature point and recording the current number of rolling cycles when a preset inflection point is identified, further includes: The continuous data of acoustic emission cumulative energy after the number of rolling revolutions corresponding to the first feature point are subjected to secondary difference processing to obtain the continuous curve of acoustic emission cumulative energy. Calculate the energy increment sequence of adjacent second preset number of cycles of the continuous curve of acoustic emission cumulative energy, perform local linear fitting on the energy increment sequence to obtain the fitting slope, and set the average slope of multiple second preset number of cycles as the second reference slope. When the fitting slope of multiple consecutive adjacent intervals of the second preset number of cycles exceeds a preset multiple of the second reference slope and shows a monotonically increasing trend, the starting number of rolling cycles of the first interval that meets the condition is set as the second feature, and the number of rolling cycles corresponding to the second feature point is recorded.

[0011] Furthermore, the step of comprehensively analyzing the multi-parameter continuous data following the specific feature point according to preset rules, and determining that the caster under test has reached a fatigue state and recording the total number of rolling revolutions when any two or more of the preset rules occur, includes: A comprehensive evaluation is performed on the continuous multi-parameter data after the number of rolling revolutions corresponding to the second feature point. When the load fluctuation continuous data falls within the preset load fluctuation threshold range, it is marked as the first preset rule case. When the radial runout value exceeds the preset radial runout safety range, it is marked as the second preset rule case. When the acoustic emission cumulative energy continuous data exceeds the preset energy safety threshold, it is marked as the third preset rule case. If the first preset rule situation and the second preset rule situation, or the first preset rule situation and the third preset rule situation, or the second preset rule situation and the third preset rule situation, or all three preset rule situations occur simultaneously, the caster under test is determined to have reached a fatigue state, and the corresponding total number of rolling revolutions is recorded.

[0012] Furthermore, before the step of determining that the caster under test has reached a fatigue state, the method further includes: Missing segments are detected in multi-parameter continuous data to obtain missing segment markers. Anomaly test results are then determined based on the missing segment markers to obtain the test validity results. If the test validity result is invalid, the test process is interrupted and an interruption execution instruction is obtained; The interrupt execution instruction prepares the spare caster for switching and a restart signal is received.

[0013] This invention also proposes a caster fatigue life testing system, comprising: The acquisition module is used to acquire the initial state reference value of the caster under test, control the caster under test to roll continuously, and acquire multi-parameter continuous data corresponding to the initial state reference value. The identification module is used to identify the multi-parameter continuous data in stages and levels. When a preset inflection point is identified, a specific feature point is set and the current number of scrolling cycles is recorded. The preset inflection point is a data point in the multi-parameter continuous data whose data change rate exceeds a preset threshold. The analysis module is used to perform comprehensive analysis on the multi-parameter continuous data after the specific feature point according to preset rules. When any two or more of the preset rules occur, the caster under test is determined to have reached a fatigue state and the total number of rolling revolutions at this time is recorded. The preset rules are set according to the multi-parameter continuous data.

[0014] The present invention also proposes a computer device comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method.

[0015] The present invention also proposes a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described method.

[0016] Compared with the prior art, this application has the following beneficial effects: This application proposes a caster fatigue life testing method and related system. By real-time acquisition of continuous data on multiple parameters such as radial load fluctuation, radial runout, and acoustic emission energy, and by adopting a phased, step-by-step identification of preset inflection points and a preset rule based on multi-parameter comprehensive analysis to determine the fatigue state, this method overcomes the shortcomings of existing technologies that rely on single macroscopic failure characteristics such as obvious deformation, bracket breakage, bearing jamming, or significant load drop, which are characterized by strong subjectivity and poor consistency. This method achieves a fundamental shift in fatigue failure criteria from endpoint macroscopic phenomena to the damage evolution process itself. It makes the determination of the final fatigue life value highly objective, quantifiable, and traceable, avoiding the phenomenon of several times the difference in life data between different testers and different laboratories, and improving the repeatability and industry comparability of test results. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] The structures, proportions, sizes, etc., shown in the accompanying drawings are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed in the specification, and are not intended to limit the implementation conditions of this application. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size should still fall within the scope of the technical content disclosed in this application, provided that they do not affect the effects and purposes that this application can produce.

[0019] Figure 1 This is a schematic diagram of the steps of a caster fatigue life test method in one embodiment of the present invention; Figure 2 This is a schematic block diagram of a caster fatigue life testing system according to an embodiment of the present invention; Figure 3 This is a schematic block diagram of the structure of a computer device according to an embodiment of the present invention; The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0021] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of features, integers, steps, operations, elements, modules, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, modules, components, and / or groups thereof. It should be understood that when an element is referred to as “connected” or “coupled” to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. Furthermore, “connected” or “coupled” as used herein may include wireless connection or wireless coupling. The term “and / or” as used herein includes all or any modules and all combinations of one or more associated listed items.

[0022] It will be understood by those skilled in the art that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. It should also be understood that terms such as those defined in general dictionaries should be understood to have the same meaning as in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as herein.

[0023] Reference Figure 1 This invention provides a method for testing the fatigue life of casters, comprising the following steps: S1: Obtain the initial state reference value of the caster under test, control the caster under test to roll continuously and obtain multi-parameter continuous data corresponding to the initial state reference value; In step S1, the caster to be tested is fixedly mounted on a general-purpose caster rolling fatigue testing bench, which consists of a large-diameter drive roller, a height-adjustable load application platform, a DC or servo motor drive system, and a simple PLC control cabinet. The caster axle is horizontally fixed on a support under the load platform, ensuring that the wheel surface is in complete contact with the surface of the roller below. A rubber layer is applied to the roller surface to simulate actual ground friction. The platform height is adjusted by a screw or cylinder to subject the caster to a constant radial load, which is set to 1.2 to 1.5 times the caster's nominal dynamic load. For example, a caster with a rated dynamic load of 500 kg is subjected to a vertical downward load of 600 to 750 kg. The load is monitored in real time by the force sensor built into the load platform to ensure that the fluctuation does not exceed ±3% throughout the process. Before the continuous rolling begins, the initial state reference value is recorded. The roller is rotated at low speed for 10 to 20 revolutions to eliminate installation gaps and uneven initial contact, and then the machine is stopped immediately. At the moment when the number of rolling revolutions N=0, the initial values ​​of multiple parameters are recorded simultaneously. These parameters include radial load F0, measured directly using a force sensor or a button-type pressure sensor mounted on the axle support and connected to an industrial control computer; the initial value is recorded as F0 = 7350 N. The second parameter is radial runout D0, measured using a non-contact laser displacement sensor aligned with the most convex point on the outer edge of the caster body; the initial runout value is recorded, for example, D0 = 0.08 mm. The third parameter is acoustic emission cumulative energy E0, measured using an acoustic emission sensor with a magnetic base near the axle on the bracket; background noise is collected for 10 seconds in a static state and then zeroed; the initial cumulative energy is recorded as E0 = 0 mV·s or directly zeroed. The fourth parameter can assist in recording initial appearance photos; an industrial camera is used to take photos of the wheel surface and bracket connection as the initial appearance benchmark for subsequent comparison of wear and cracks. After recording the initial benchmark values, the drive roller is started, causing the caster to continuously roll at a common test speed of 2~3 m / s (corresponding to a roller speed of approximately 80~120 r / min), and the multi-parameter continuous data acquisition phase officially begins from N=0 revolutions. With each rotation, the PLC or industrial computer automatically increments the rotation count N, simultaneously recording the current radial load F(N), radial runout D(N), and the cumulative energy E(N) of the acoustic emission signal after rectification. These three signals correspond to the rotation count N, forming three curves that continuously change with the rotation count. Every 100 or 200 rotations, the acoustic emission event count and average energy for that interval are calculated and saved. To prevent misjudgment by a single sensor, an industrial camera is automatically triggered by a relay every 5000 rotations, and all photo filenames automatically include the current rotation count and timestamp.

[0024] S2: The multi-parameter continuous data is identified in stages and levels. When a preset inflection point is identified, a specific feature point is set and the current number of rolling cycles is recorded. The preset inflection point is a data point in the multi-parameter continuous data whose data change rate exceeds a preset threshold. In step S2, the radial runout D(N) in step S1 is the peak-to-peak radial runout of the wheel during rolling, and the maximum runout value is recorded for each revolution. In one example, the phased identification can be divided into two stages. The first stage identifies the initial damage point (the first preset inflection point), and the D(N) curve is smoothed by a moving average of 5000 to 10000 revolutions to filter out normal rolling noise. Then, the local slope K is calculated every 10000 revolutions. i (A simple fit can be achieved using the least squares method). Let the average slope of the first 100,000 laps after the start of the experiment be K0 (usually close to 0). When K0 is within three consecutive 10,000-lap intervals... i When all values ​​are greater than 1.5*K0, and the slope itself shows a monotonically increasing trend, the caster is considered to have left the initial break-in period and entered a stage of slow damage accumulation. At this point, the starting number of revolutions in the first interval meeting the conditions is recorded as N1, i.e., the first specific characteristic point. For example, in a certain test, for the first 80,000 revolutions, D(N) was basically stable at 0.35mm, and the slope K0≈0.00002mm / revolution. From the 120,000th revolution onwards, the slopes of three consecutive intervals reached 0.00004, 0.00005, and 0.00006mm / revolution respectively, all exceeding 1.5 times K0 and continuously increasing. Therefore, N1=120,000 revolutions. In the second stage, after N≥N1, the energy increment ΔE of E(N) is taken every 5,000 revolutions, and then a local linear fit is performed on the ΔE sequence to obtain the slope S. j Take the average slope S of the first 50 units (i.e., the 250,000 cycles after N1). avg As a benchmark. When the S of 5 consecutive units j A sudden increase, and any one of the S's j ≥2.0×S avg When the crack enters the accelerated propagation phase, the starting point of the first element exceeding the threshold is denoted as N2. For example, in actual measurements, after N1, the acoustic emission energy increases slowly, averaging about 800 mV·ms every 5000 cycles. Starting from the 480,000th cycle, the energy increment slope of five consecutive elements suddenly jumps to 2.8 times the original value, accompanied by several high-energy events greater than 5000 mV·ms. In this case, N2 = 480,000 cycles.

[0025] S3: Perform comprehensive analysis on the multi-parameter continuous data after the specific feature point. When a preset condition is detected, determine that the caster under test has reached a fatigue state and record the total number of rolling revolutions at this time. The preset condition is that the value of a specific parameter in the multi-parameter continuous data exceeds a preset fatigue life determination threshold. In step S3, the specific feature points refer to the multiple preset inflection points identified in step S2, namely N1 (damage initiation point) and N2 (accelerated propagation start point). Rapid instability judgment begins from the stage after N2. At this time, the internal damage of the caster has entered the stage of rapid crack propagation from slow accumulation, and the material properties are about to undergo irreversible overall instability. Therefore, a multi-parameter joint rule is used to capture the moment of impending functional failure. In this embodiment, the preset rule includes the abrupt change in the slope of the radial runout D(N). After entering the N2 stage, the local slope is calculated every 2000 revolutions. If the slope of the current 2000 revolution interval reaches more than 2.5 times the average slope of the previous stable period, it is considered that the bearing cage has become severely loose or the balls have pitted and peeled off, resulting in a sharp increase in runout, and this rule is triggered. The second case is the explosive growth of the instantaneous event rate of the acoustic emission signal. The number of acoustic emission events (peaks with amplitude exceeding a set threshold) is counted every revolution. When the single-revolution event rate first exceeds 8 times the average event rate of the entire test history, this rule is triggered. For example, in the early stages of the test, there were only 3-5 events per lap on average. However, during the rapid instability phase, 42 events suddenly occurred in a single lap, far exceeding the threshold by 8 times. This indicates that internal cracks are expanding and forming macroscopic cracks. The third scenario is a continuous abnormal drop in radial load. Under normal circumstances, the load fluctuation does not exceed ±3%. However, when the caster bracket or bearing is severely deformed, causing a decrease in the effective support height, the force sensor will detect a continuous load drop of more than 8% that cannot be recovered. For example, in one test, when the caster reached 680,000 laps, the load dropped continuously from the set 2000N to below 1780N and remained below 1780N for more than 8000 laps without recovering. This directly indicates that the structure has undergone permanent deformation. The fourth scenario is a clearly visible change in appearance damage. A high-resolution photo is automatically taken every 5000 laps using an industrial camera. The photo sequence within the last 10000 laps is examined visually or using free and open-source image recognition software (such as calling a pre-trained crack detection model). Once a visible crack is found in the bracket, obvious elliptical deformation of the wheel body, or the bearing seal comes loose, this rule is triggered. For example, in one test, a 5mm long through crack was clearly visible in the photo taken after 730,000 revolutions, while the previous photo (at 680,000 revolutions) showed the caster was completely normal, thus triggering this rule. When any two or more of the above four conditions occur almost simultaneously within a short window of ±5,000 revolutions (approximately less than half an hour in actual time), the caster is determined to have entered an irreversible rapid instability failure stage. The recorded number of rolling revolutions N3 at this time is the final fatigue life value L of the caster. For example, in a real test, when the caster reached 718,000 revolutions, it simultaneously met the following three rules: ① radial runout slope exceeds 2.8 times, ② acoustic emission single-cycle event rate reaches 11 times the historical average, and ③ the appearance photo shows cracks in the support. An alarm was immediately triggered and the rollers were stopped, recording L=718,000 revolutions.Even if a sensor occasionally experiences noise interference, the test can be reliably terminated as long as two other independent parameters corroborate this. Conversely, if only a single parameter is abnormal, the test continues, thus avoiding overestimation of lifespan due to misjudgment.

[0026] It is worth noting that the equipment and sensors in the above steps can all be easily implemented with existing technology. For example, a general-purpose caster rolling fatigue testing bench can be customized with different specifications and sizes to adapt to the testing of various casters according to actual needs. The connection and data transmission between these devices can be achieved by transmitting the data collected by the sensors to a PLC or industrial control computer for storage and analysis via wired or wireless means.

[0027] In summary, by organically combining continuous monitoring of multiple parameters, phased inflection point identification, and comprehensive judgment based on multiple rules, the fatigue life of casters has been transformed from rough experience-based judgment to precise data-based judgment. It has many advantages, such as strong objectivity, high repeatability, outstanding scientific rigor, low implementation cost, good safety, and the ability to provide a complete data chain for reliability analysis. It solves the industry problems of vague fatigue endpoints, poor comparability, and strong subjectivity in existing technologies, and has significant innovation, outstanding substantive features, and significant progress.

[0028] In another embodiment, the aforementioned multi-parameter continuous data can also be acquired using other data acquisition methods or different types of sensors, as long as they can accurately reflect the state changes of the caster during rolling. For example, in addition to the parameters mentioned above such as radial load, radial runout, and cumulative acoustic emission energy, a temperature sensor can be added to monitor temperature changes during caster rolling, as abnormal temperature increases may indicate damage or failure of the caster's internal structure. Furthermore, the application can be extended to fatigue life testing of different types of casters or similar rolling components; simply adjust the test parameters and judgment rules according to the specific characteristics of the object.

[0029] In another test embodiment, referring to Table 1, the evolution process of multi-parameter continuous data and the results of key feature point extraction in a single test are shown. The test object was an industrial caster with a rated dynamic load of 500 kg, with an applied radial load of approximately 750 kg (7350 N, approximately 1.47 times the nominal dynamic load) and a roller linear speed of 2.5 m / s. Initially, the radial runout D0 was only 0.08 mm, and after the acoustic emission accumulated energy was cleared, E0 = 0 mV·s, and the load stabilized at 7350 N. During the first 120,000 rolls, the caster was in a normal break-in period, and the radial runout D(N) was basically stable at around 0.35 mm, the acoustic emission energy accumulated slowly, and the load fluctuation was controlled within ±3%. When the 120,000th roll was reached, the slope K of the D(N) curve, calculated by sliding average and local slope, was calculated for three consecutive 10,000-roll intervals. iAll values ​​exceeded 1.5 times the initial baseline slope K0 and showed a monotonically increasing trend, indicating the initial appearance of damage. The first specific characteristic point N1 = 120,000 cycles was recorded, at which point the runout had slowly increased to approximately 0.42 mm. The experiment continued, and at the 480,000th cycle, the local slope Sj of the acoustic emission energy increment ΔE suddenly exceeded 2.0 times the previous average slope Savg for five consecutive 5,000-cycle units. At the same time, there were multiple high-energy releases with single events exceeding 5,000 mV·ms, indicating that the crack had entered the accelerated propagation stage. The second specific characteristic point N2 = 480,000 cycles was recorded, at which point the cumulative energy had reached 3.12 million mV·s, and the runout had increased to 0.68 mm. Upon entering the N2 stage, monitoring of rapid instability was strengthened. When the caster reached 718,000 revolutions, three independent judgment rules were triggered simultaneously within a short window of 5,000 revolutions: the radial runout reached 2.8 times the local slope of the previous stage within 2,000 revolutions; the single-revolution acoustic emission event rate suddenly increased to 11 times the historical average; and an industrial camera clearly showed a through crack of about 6mm in length on the support base plate. At the same time, the force sensor detected that the load continued to drop to 6420N (a drop of about 12.7% that could not be recovered). Since any two or more rules were met simultaneously in a very short time, it was determined that the caster had entered the irreversible functional failure stage, and an automatic alarm was triggered to stop the rollers. The fatigue life of the caster was ultimately recorded as L=718,000 revolutions. This table visually illustrates the complete technical solution of this embodiment, which identifies inflection points N1 and N2 in stages and levels by using multiple parameters (radial runout, acoustic emission, load, and appearance), and then accurately captures the endpoint N3 by integrating rapid instability rules. This avoids underestimation of life or excessive damage caused by a single indicator (such as only looking at the runout exceeding 3mm or only achieving complete structural fracture), significantly improves the objectivity, repeatability, and safety of test results, and greatly shortens the invalid and excessively long test time, thus having significant industrial application value.

[0030] Table 1: In one embodiment, the step of obtaining the initial state reference value of the caster under test includes: The initial state reference value is obtained by acquiring the force signal, radial runout and yaw amplitude and peak amplitude of the caster under test based on the force sensor, displacement sensor and acoustic emission sensor on the caster under test.

[0031] In the above embodiment, when the caster under test is horizontally fixed below the liftable load platform, with the wheel surface and drive roller fully in contact and 1.2 to 1.5 times the rated dynamic load applied, the roller is first allowed to rotate at low speed for 10 revolutions to fully eliminate installation gaps and uneven initial friction. At this time, the caster is in a stable initial state that is closest to the actual working condition. The stable readings of three sensors are read simultaneously, including the current actual radial force signal output by the force sensor placed at the wheel axle support, which serves as the initial load reference F0, typically fluctuating less than ±2N; the non-contact laser displacement sensor measures the outer surface of the wheel multiple times, and the mean and standard deviation of multiple samples are taken to obtain the initial radial runout and sway amplitude D0 (typically less than 0.05mm) and the initial sway amplitude; the acoustic emission sensor, which is close to the bracket near the wheel axle, collects background noise for more than 10 seconds in a quiet environment, and then the cumulative acoustic emission energy at this time is cleared or the extremely low initial energy value E0 is recorded, while the amplitude peak distribution is recorded as the initial acoustic characteristics. These initial values ​​are uniformly labeled as the baseline parameters when the number of rolling revolutions N=0, and saved as the zeroing reference for all subsequent continuous data. An industrial camera also captures an appearance photograph at this time as an initial visual baseline. The initial state baseline values ​​obtained in this way are repeatable, avoiding systematic errors caused by installation differences or sensor drift, and providing a reliable comparative basis for subsequent identification of minor damage accumulation. The entire process is usually completed within 30 seconds before the formal high-speed rolling test begins.

[0032] In one embodiment, the step of controlling the caster under test to continuously roll and acquiring multi-parameter continuous data corresponding to the initial state reference value includes: The caster under test is controlled to roll continuously, and the number of rolls is obtained from the photoelectric sensor inside the caster under test. Obtain continuous data corresponding to the number of rolling revolutions of the initial state reference value, and divide the continuous data into data segments according to the preset number of revolutions; Based on each data segment, continuous load fluctuation data is calculated based on the triaxial force signal, continuous radial runout data is calculated based on the radial runout and yaw amplitude, and continuous acoustic emission cumulative energy data is calculated based on the yaw amplitude and peak amplitude, thus obtaining multi-parameter continuous data.

[0033] In the above embodiment, the constant linear velocity of the drive roller causes the caster to roll continuously. A high-precision photoelectric encoder installed inside the wheel axle or at the end of the roller axle outputs a pulse signal for each revolution. The industrial control computer accumulates the pulse count in real time to obtain the precise number of rolling revolutions N, and uses N as a unified time axis. Three raw signals are synchronously acquired at a sampling frequency of not less than 1000Hz, including the radial force F(t) output by the force sensor in real time, the radial displacement D(t) of the wheel body output by the laser displacement sensor, and the raw waveform of the elastic wave output by the acoustic emission sensor. Every 100 revolutions is considered a data segment, triggering a centralized processing. Within these 100 revolutions, the three-axis force signals are statistically analyzed to extract the mean, standard deviation, and maximum fluctuation of the load fluctuation amplitude, thereby generating a continuous load fluctuation data sequence F. var (N); Envelope detection and low-pass filtering are performed on the raw radial runout data to obtain the peak-to-peak value and yaw amplitude of radial runout every 100 revolutions, forming continuous radial runout data D. pp (N); Perform event extraction and energy integration on the acoustic emission signal, calculate the cumulative energy increment of newly generated acoustic emission within the 100 cycles, and add it to the total energy to form continuous data of cumulative acoustic emission energy E. cum (N). Three multi-parameter continuous data curves are obtained, with the number of rolling revolutions N as the abscissa and corresponding to the initial baseline value. At the same time, an industrial camera is automatically triggered to take pictures every 5000 revolutions, forming a discrete but precisely corresponding appearance image sequence, which forms the original multi-parameter dataset.

[0034] In one embodiment, the step of performing staged, step-by-step identification of the multi-parameter continuous data includes: A sliding window with a preset number of bits is constructed to smooth the continuous radial runout data, resulting in a smoothed radial runout data curve. The local slope of the radial runout data curve is calculated according to the first preset number of cycles interval, and the average slope of multiple first preset number of cycles intervals is set as the first reference slope. When the local slopes of multiple consecutive adjacent intervals of the first preset number of cycles all exceed a preset multiple of the first reference slope and show a monotonically increasing trend, the starting number of rolling cycles of the first interval that meets the condition is set as the first feature point, and the number of rolling cycles corresponding to the first feature point is recorded.

[0035] In the above embodiment, the radial runout continuous data D pp (N) A median or low-pass filter is used to smooth the curve using a sliding window with a width of 5000–10000 cycles to eliminate high-frequency noise and obtain a smooth radial jump trend curve. The curve is then segmented according to a first preset number of cycles (which can be 10000 cycles), and the local linear regression slope K is calculated once for each interval. iThe average slope K0 of the first 10-20 intervals is calculated as the first reference slope, representing the normal slow growth rate during the initial break-in period. K0 is then monitored in real-time for each subsequent interval. i When a local slope K is found in 3 to 5 consecutive adjacent intervals i All of them exceed 1.5 to 2.0 times the first reference slope K0, and these slopes themselves show a monotonically increasing trend (i.e., K0 > 100%). {i+1} >K i When the caster has passed its break-in period and irreversible micro-damage begins to appear internally, it is determined that the caster has exited the stable break-in period. At this point, the initial number of rolling revolutions in the first interval that meets the above two conditions is set as the first characteristic point N1 and recorded. This judgment logic is based on the physical nature of radial runout, which changes from slow linear growth to accelerated growth. In the early stages, it is mainly caused by surface wear, with a small and stable slope. When micro-cracks or plastic deformation occur in the bearing cage or raceway, the radial runout will suddenly accelerate and form a clear inflection point. By combining the judgment with three conditions—multiple consecutive intervals, exceeding a preset multiple, and monotonically increasing—false judgments caused by single noise disturbances are effectively avoided. This makes the identification of the first characteristic point highly robust and objective, and it usually appears in the 20% to 40% range of the total lifespan, providing a clear dividing point for subsequent key monitoring of the rapid deterioration stage.

[0036] In one embodiment, the step of performing phased and progressive identification of the multi-parameter continuous data, setting a specific feature point and recording the current number of scroll cycles when a preset inflection point is identified, further includes: The continuous data of acoustic emission cumulative energy after the number of rolling revolutions corresponding to the first feature point are subjected to secondary difference processing to obtain the continuous curve of acoustic emission cumulative energy. Calculate the energy increment sequence of adjacent second preset number of cycles of the continuous curve of acoustic emission cumulative energy, perform local linear fitting on the energy increment sequence to obtain the fitting slope, and set the average slope of multiple second preset number of cycles as the second reference slope. When the fitting slope of multiple consecutive adjacent intervals of the second preset number of cycles exceeds a preset multiple of the second reference slope and shows a monotonically increasing trend, the starting number of rolling cycles of the first interval that meets the condition is set as the second feature, and the number of rolling cycles corresponding to the second feature point is recorded.

[0037] In the above embodiment, when the number of rolling revolutions N exceeds the first feature point N1, the cumulative acoustic emission energy E after this segment is calculated. cum (N) Perform a second difference process. First, calculate the first difference to obtain the energy increment sequence ΔE(N). Then, perform a second difference on ΔE(N) or directly perform local linear fitting on the ΔE sequence according to the second preset number of cycles (usually 5000 cycles) to obtain the fitting slope S of each interval. jThis slope reflects the acceleration of the crack propagation energy release rate. The average fitted slope S for the first 50 intervals was also calculated. avg As a second reference slope, S is continuously monitored in subsequent intervals. j When the fitted slope of 4 to 6 consecutive adjacent intervals exceeds 2.0 to 3.0 times the second reference slope S avg When these slopes themselves exhibit a clear monotonically increasing trend, it is determined that the internal crack has changed from slow and stable propagation to accelerated and unstable propagation, accompanied by a concentrated outbreak of high-energy acoustic emission events. At this point, the initial rolling number of the first interval that meets the conditions is set as the second characteristic point N2 and recorded. This embodiment utilizes the sensitivity of acoustic emission technology to the instantaneous propagation of microcracks, i.e., in the early damage stage, acoustic emission events are sparse, and the energy release rate is low and stable. When the crack length exceeds the critical value, the propagation rate increases sharply, and the energy release exhibits an intensified characteristic. Identification is performed through quadratic difference or local fitting slope abrupt change, which is more earlier and more accurate than simply looking at the cumulative energy threshold. Typically, N2 appears at 60% to 80% of the total lifespan, complementing the macroscopic manifestation of radial runout, further subdividing the fatigue process into three clear stages: the break-in period, the slow damage period, and the accelerated propagation period, laying the foundation for the second inflection point in determining the final lifespan end.

[0038] In one embodiment, the step of comprehensively analyzing the multi-parameter continuous data after the specific feature point, and determining that the caster under test has reached a fatigue state and recording the total number of rolling revolutions at this time when a preset condition is met includes: A comprehensive evaluation is performed on the continuous multi-parameter data after the number of rolling revolutions corresponding to the second feature point. When the load fluctuation continuous data falls within the preset load fluctuation threshold range, it is marked as the first preset rule case. When the radial runout value exceeds the preset radial runout safety range, it is marked as the second preset rule case. When the acoustic emission cumulative energy continuous data exceeds the preset energy safety threshold, it is marked as the third preset rule case. If at least two of the above-mentioned first, second, and third preset rule conditions occur simultaneously, the caster under test is determined to have reached a fatigue state, and the total number of rolling revolutions at this time is recorded.

[0039] In the above embodiment, when the number of rolling cycles N > N2, the three curves are simultaneously evaluated in real time for both threshold and trend. The first preset rule is for continuous load fluctuation data. When the load standard deviation or peak-to-peak value in the most recent 5000 to 10000 cycles continuously exceeds the initial baseline by 1.5 to 2 times, or when there is a continuous load drop of ≥8%, it is marked as load instability. The second preset rule is for continuous radial runout data. When the smoothed radial runout value continuously exceeds the initial value by 3 to 5 times the safety range (e.g., the initial safety upper limit of 0.05 mm is set to 0.2 to 0.3 mm), it is marked as runout out of control. The third preset rule is for cumulative acoustic emission energy. When the energy increment in the most recent 5000 cycles exceeds 10 times the historical average, or when the cumulative energy curve shows an approximately exponential surge, it is marked as a high-energy event cluster. The system employs a 2-out-of-3 or all three conditions must be met logic. This means that if any two or three rules are triggered simultaneously within a narrow window of ±3000 to 5000 rotations, the caster is determined to have entered an irreversible, rapid instability and failure phase. At this point, the current number of rotations is recorded as the total fatigue life L=N. fatigue This multi-parameter redundancy judgment mechanism improves the objectivity and anti-interference ability of the results. A single parameter may exceed the limit due to accidental factors, but the three parameters come from completely different physical mechanisms (force, displacement, and acoustics), and are almost unlikely to be falsely triggered at the same time. In the event of real fatigue instability, the three parameters will inevitably show highly correlated abrupt changes. Through this strict rule that any two or more parameters are satisfied at the same time, the determination of the final life end no longer depends on subjective visual fracture, but achieves a completely data-driven, automatically executed, and repeatable objective standard. This point usually occurs 1000 to 3000 revolutions before the bearing completely seizes or the bracket breaks, perfectly realizing the determination of fatigue life rather than destructive life.

[0040] In one embodiment, before the step of determining that the caster under test has reached a fatigue state, the method further includes: Missing segments are detected in multi-parameter continuous data to obtain missing segment markers. Anomaly test results are then determined based on the missing segment markers to obtain the test validity results. If the test validity result is invalid, the test process is interrupted and an interruption execution instruction is obtained; The interrupt execution instruction prepares the spare caster for switching and a restart signal is received.

[0041] In the above embodiments, continuous data on load fluctuation, radial runout, and acoustic emission cumulative energy are checked in real time throughout the test. When any parameter's continuous data is missing and the missing length exceeds a preset threshold for the number of missing cycles, the rolling cycle interval containing the missing segment is marked, and the starting and ending cycles of the missing segment are recorded, resulting in multi-parameter continuous data with missing segment markings. Based on the missing segment markings, all marked missing segments are cumulatively counted. When the cumulative number of missing cycles or the number of missing cycles in a single segment for any parameter exceeds a preset maximum tolerable number of missing cycles, the current test is determined to be an abnormal test, and a test invalidity conclusion is output. Simultaneously, the current test process is automatically terminated, and the rolling cycle count at termination is recorded, resulting in test status data containing the test validity conclusion. When an abnormal test is determined and a test invalidity conclusion is obtained, an interrupt execution command is sent to the main control unit of the test system based on the conclusion. This causes the test bench to immediately stop applying rolling load to the caster under test and maintain its current mechanical position locked. At the same time, all sensor data acquisition channels are shut down, resulting in a test status that has been interrupted. After the interruption command is completed, an invalid test prompt will automatically pop up and the number of the currently failed caster and the number of rotations will be recorded. At the same time, the spare casters of the same batch will be fixed on the platform for testing. When it is detected that the spare casters are correctly installed and the initial state reference value is re-acquired, a restart signal will be generated and the rotation count counter will be cleared, so that the test system can restart all test steps.

[0042] Reference Figure 2 A caster fatigue life testing system, comprising: The acquisition module 100 is used to acquire the initial state reference value of the caster under test, control the caster under test to roll continuously, and acquire multi-parameter continuous data corresponding to the initial state reference value. The identification module 200 is used to identify the multi-parameter continuous data in stages and levels. When a preset inflection point is identified, a specific feature point is set and the current number of scrolling cycles is recorded. The preset inflection point is a data point in the multi-parameter continuous data whose data change rate exceeds a preset threshold. The analysis module 300 is used to perform comprehensive analysis on the multi-parameter continuous data after the specific feature point. When a preset condition is detected, the caster under test is determined to have reached a fatigue state and the total number of rolling revolutions at this time is recorded. The preset condition is that the value of a specific parameter in the multi-parameter continuous data exceeds a preset fatigue life determination threshold.

[0043] Reference Figure 3 This application also provides a computer device, which may be a server, and its internal structure may be as follows: Figure 3As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores database data. The network interface allows communication with external terminals via a network connection. When executed by the processor, the computer program implements a caster fatigue life testing method.

[0044] One embodiment of this application also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements a method for testing the fatigue life of a caster, including the following steps: acquiring an initial state reference value of the caster under test; controlling the caster under test to continuously roll and acquiring multi-parameter continuous data corresponding to the initial state reference value; performing stage-by-stage identification on the multi-parameter continuous data; setting a specific feature point and recording the current number of rolling revolutions when a preset inflection point is identified; wherein the preset inflection point is a data point in the multi-parameter continuous data where the rate of change of data exceeds a preset threshold; performing comprehensive analysis on the multi-parameter continuous data after the specific feature point; and determining that the caster under test has reached a fatigue state and recording the total number of rolling revolutions at this time when a preset condition is detected; wherein the preset condition is that a specific parameter value in the multi-parameter continuous data exceeds a preset fatigue life determination threshold.

[0045] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media provided in this application and used in the embodiments can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual-speed SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0046] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A method for testing the fatigue life of casters, characterized in that, include: Obtain the initial state reference value of the caster under test, control the caster under test to roll continuously and obtain multi-parameter continuous data corresponding to the initial state reference value; The multi-parameter continuous data is identified in stages and levels. When a preset inflection point is identified, a specific feature point is set and the current number of scrolling cycles is recorded. The preset inflection point is a data point in the multi-parameter continuous data whose data change rate exceeds a preset threshold. A comprehensive analysis is performed on the continuous multi-parameter data following the specific feature point. When a preset condition is detected, the caster under test is determined to have reached a fatigue state, and the total number of rolling revolutions at this time is recorded. The preset condition is that the value of a specific parameter in the continuous multi-parameter data exceeds a preset fatigue life determination threshold.

2. The method for testing the fatigue life of a caster according to claim 1, characterized in that, The step of obtaining the initial state reference value of the caster under test includes: The initial state reference value is obtained by acquiring the force signal, radial runout and yaw amplitude and peak amplitude of the caster under test based on the force sensor, displacement sensor and acoustic emission sensor on the caster under test.

3. The method for testing the fatigue life of a caster according to claim 2, characterized in that, The step of controlling the caster under test to roll continuously and acquiring multi-parameter continuous data corresponding to the initial state reference value includes: The caster under test is controlled to roll continuously, and the number of rolls is obtained from the photoelectric sensor inside the caster under test. Obtain continuous data corresponding to the number of rolling revolutions of the initial state reference value, and divide the continuous data into data segments according to the preset number of revolutions; Based on each data segment, continuous load fluctuation data is calculated based on the triaxial force signal, continuous radial runout data is calculated based on the radial runout and yaw amplitude, and continuous acoustic emission cumulative energy data is calculated based on the yaw amplitude and peak amplitude, thus obtaining multi-parameter continuous data.

4. The method for testing the fatigue life of a caster according to claim 3, characterized in that, The step of performing phased and progressive identification of the multi-parameter continuous data includes: A sliding window with a preset number of bits is constructed to smooth the continuous radial runout data, resulting in a smoothed radial runout data curve. The local slope of the radial runout data curve is calculated according to the first preset number of cycles interval, and the average slope of multiple first preset number of cycles intervals is set as the first reference slope. When the local slopes of multiple consecutive adjacent intervals of the first preset number of cycles all exceed a preset multiple of the first reference slope and show a monotonically increasing trend, the starting number of rolling cycles of the first interval that meets the condition is set as the first feature point, and the number of rolling cycles corresponding to the first feature point is recorded.

5. The method for testing the fatigue life of a caster according to claim 4, characterized in that, The step of performing phased and progressive identification of the multi-parameter continuous data, setting a specific feature point and recording the current number of scroll cycles when a preset inflection point is identified, further includes: The continuous data of acoustic emission cumulative energy after the number of rolling revolutions corresponding to the first feature point are subjected to secondary difference processing to obtain the continuous curve of acoustic emission cumulative energy. Calculate the energy increment sequence of adjacent second preset number of cycles of the continuous curve of acoustic emission cumulative energy, perform local linear fitting on the energy increment sequence to obtain the fitting slope, and set the average slope of multiple second preset number of cycles as the second reference slope. When the fitting slope of multiple consecutive adjacent intervals of the second preset number of cycles exceeds a preset multiple of the second reference slope and shows a monotonically increasing trend, the starting number of rolling cycles of the first interval that meets the condition is set as the second feature, and the number of rolling cycles corresponding to the second feature point is recorded.

6. The method for testing the fatigue life of a caster according to claim 5, characterized in that, The step of comprehensively analyzing the multi-parameter continuous data after the specific feature point, and determining that the caster under test has reached a fatigue state and recording the total number of rolling revolutions at this time when a preset condition is met, includes: A comprehensive evaluation is performed on the continuous multi-parameter data after the number of rolling revolutions corresponding to the second feature point. When the load fluctuation continuous data falls within the preset load fluctuation threshold range, it is marked as the first preset rule case. When the radial runout value exceeds the preset radial runout safety range, it is marked as the second preset rule case. When the acoustic emission cumulative energy continuous data exceeds the preset energy safety threshold, it is marked as the third preset rule case. If at least two of the above-mentioned first, second, and third preset rule conditions occur simultaneously, the caster under test is determined to have reached a fatigue state, and the total number of rolling revolutions at this time is recorded.

7. The method for testing the fatigue life of a caster according to claim 1, characterized in that, Before the step of determining that the caster under test has reached a fatigue state, the method further includes: Missing segments are detected in multi-parameter continuous data to obtain missing segment markers. Anomaly test results are then determined based on the missing segment markers to obtain the test validity results. If the test validity result is invalid, the test process is interrupted and an interruption execution instruction is obtained; The interrupt execution instruction prepares the spare caster for switching and a restart signal is received.

8. A caster fatigue life testing system, characterized in that, include: The acquisition module is used to acquire the initial state reference value of the caster under test, control the caster under test to roll continuously, and acquire multi-parameter continuous data corresponding to the initial state reference value. The identification module is used to identify the multi-parameter continuous data in stages and levels. When a preset inflection point is identified, a specific feature point is set and the current number of scrolling cycles is recorded. The preset inflection point is a data point in the multi-parameter continuous data whose data change rate exceeds a preset threshold. The analysis module is used to perform comprehensive analysis on the continuous multi-parameter data after the specific feature point. When a preset condition is detected, the caster under test is determined to have reached a fatigue state and the total number of rolling revolutions at this time is recorded. The preset condition is that the value of a specific parameter in the continuous multi-parameter data exceeds a preset fatigue life determination threshold.

9. A computer device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 7.

Images