LOCATION OF FAULTS OCCURRING DURING HIGH-SPEED TESTS

Abstract

The invention is a fault and location detection system and method (100) for detecting and locating faults in indoor high-speed tests, and includes the following steps: Detection of at least one barcode (8) location on the tire (3) by means of a differentiable binary equalization based detection model; Obtaining a sequence of boxes (10) by time sampling method by referencing the detected barcode (8) location; Obtaining a spatial feature array by processing the mentioned box array (10) by convolutional neural network method (CNN); Obtaining a time-dependent feature array by processing the spatial feature array by liquid neural network (LTC / LNN); Obtaining the classification result by processing the mentioned time-dependent feature array in multi-head attention block (MHA).

Description

1 TARIFF LOCATION OF FAULTS OCCURRING DURING HIGH-SPEED TESTS Technical Area The invention relates to the location of faults that occur during high-speed testing. The invention specifically enables the detection and location of faults during high-speed indoor tests. It involves a fault and location detection system and method for identification purposes. State of the Art Today, in order to evaluate the high-speed durability of tires... Tests conducted under laboratory conditions have long been used in the automotive industry. It is being implemented. One of the known applications in this context is "high speed resistance 10". These are closed-space tests referred to as "high-speed failure tests" or "enclosed-space failure tests". In these tests, the tire is filled with air until it reaches a certain internal pressure value. It is inflated and mounted on an intermediate apparatus that moves over a steel drum. Test setups typically use drums with a diameter of 1.7–2.0 meters. It is used and loads are applied to the tire at the angles specified in the test methods. The test, 15 It starts with the tire's preheating phase; during this phase, the tire stays at moderate speeds for a certain period of time. It is brought to operating temperature by being run. In the next stage, the tire rotation speed The methods are adjusted according to the specified levels and tested over the determined periods. These tests are conducted according to international standards such as ECE R30 and FMVSS 139. The selection and decision-making process for tire production formulas is defined and determined by the manufacturers. 20 This stage is used in quality control processes and certification processes. The tests... Its purpose is to prevent structural deterioration of the tire, belt- This can manifest as types of failure such as carcass separation, sudden pressure loss, or excessive deformation. to remove, determine the tire's lifespan, and thus test the tire's reliability. This is to... In known applications, there are also some limitations to such tests. 25 Drum-based systems cannot accurately represent real road conditions, especially. Road roughness and contact area vary. Also, at high speeds, the tire... Accurate measurement of the resulting heat buildup can be limited due to sensor response times. This situation can affect the accuracy of the test results. In this context, in high-speed endurance tests of tires, the accuracy of the test and 30 To ensure repeatability, the position of the tire on the drum and its rotation 2 Monitoring its behavior is important. In this context, in known test systems, the bubble Locating a tire failure that starts and progresses in a similar manner is done by a human operator. by means of a ruler that requires calibration at the end of the test on the tire The location of the bubble on the tire will be determined according to the barcode. This is done using existing systems where the tire is mounted on a metal drum. It is being manipulated and rotated around a fixed axis throughout the testing period. Some known systems use optical encoders and magnetic sensors to track the rotational position of the tire. or is monitored via laser-based distance measuring sensors. These sensors, Reference marks on the drum or location markings placed on the tire surface It generates information about rotational speed and angular position by detecting the markers. 10 In advanced systems, lateral run-out and radial movement of the tire on the drum are considered. Radial run-out is also measured, and for this purpose, non-contact laser sensors or High-speed camera systems are used. These measurements are taken while the tire is under the rim during testing. It aims to determine the amount of displacement relative to the center. However, known test setups have some limitations in determining location. 15 Since the vibration amplitude of the tire increases especially at high speeds, the information received from the sensors... Signals can become clogged with interference, reducing measurement accuracy. Also, the test... An increase in ambient temperature can affect the measurement stability of the sensors, which can lead to testing. This can lead to discrepancies in the data. In addition, in current systems, the tire's performance... dynamic behavior on the drum (e.g., axial 20 at the microscopic level) (oscillations or lateral misalignments) can usually be monitored in real time. It is not. In conclusion, failure location detection is fundamental in current high-speed endurance testing systems. as a human operator using a ruler that requires calibration at the end of the test It is based on the principle of measurement, and this process requires time and an additional tool. 25 It has a disadvantage because it is easily audible. As a result of the research conducted on this subject, the smart active tire with the number WO2013114388A1 was identified. A PCT application titled "Tire Pressure Optimization System" was found. The application mentions tires. Pressure Optimization System (TPOS) (102) and Tire Pressure Monitoring Units (TPCU) (104) Intelligent Active Tire Pressure Optimization System (SATPOS). SATPOS, 30 especially in emergency and unavoidable critical driving situations, it instantly detects tire pressure and controlling, reducing emergency and high-speed braking distance, preventing loss of traction, in water 3 skidding, tipping over or loss of stability, excessive or insufficient steering, or a puncture. To reduce loss of control, intelligent sensing performs context-aware computing. The application does not include fault location determination in a high-speed test environment. In conclusion, due to the requirements described above and the limitations of the current solutions, Due to its shortcomings, an improvement in the relevant technical field is necessary. 5 It has been made. Purpose of the Invention The invention was created by drawing inspiration from existing situations and overcoming the aforementioned drawbacks. It aims to solve the problem. The main objective of the invention is to provide a differentiable binary equalization-based detection system on the tire. Detecting at least one barcode location through the model, the detected barcode a sampling method based on time and referencing location array of boxes, the aforementioned array of boxes using the convolutional neural network method (CNN) Obtaining the spatial feature array through processing, the liquid spatial feature array. Time-dependent feature array processing using (liquid) neural networks (LTC / LNN) method 15 The acquisition of the sequence is in the multi-headed attention block of the aforementioned time-dependent attribute sequence. (MHA) processing involves the steps involved in obtaining the classification result. Thus, in high-speed tests, the location of the fault is determined using a deep learning-based model. This is determined by the decision-making mechanism. Another objective of the invention is to create a temporally sampled sequence of boxes within a predetermined range of 20. the step of obtaining the spatial array of boxes formed by shifting the scale It includes. Another objective of the invention is to align the spatial array of boxes with the center of the rubber band. This includes ensuring that it has been provided in at least one defined search trajectory. Another objective of the invention is to radially track the aforementioned search trajectory along the tire radius of 25. The process involves providing multiple failure trajectories by shifting the scale. It includes the step. Another objective of the invention is to follow the aforementioned fault trajectory of the spatial array of boxes. shifted so that there is a predetermined distance between their centers and these 4 a number of searches were conducted in the trajectory that were possible given the distance relative to the circumference of the tire. It includes the step of being. Another purpose of the invention is to identify the fault window number and the spatial box. the predetermined slip distance of the series is found in relation to the tire circumference. The location of the fault is determined by multiplying the angle of change by 5, depending on the barcode location. It includes the detection phase. Another objective of the invention is to create at least one imaging device associated with at least one light source. the unit, at least one processor that processes the images received from the aforementioned imaging unit. The unit, mentioned as a processing unit, uses deep learning-based technology to detect tire defects. It must include at least one model-supported decision-making mechanism. Thus, in high-speed tests, 10 Monitoring the failure and location of tires while testing them under load and speed. is provided. The structural and characteristic features and all the advantages of the invention are given in the figures below. This becomes clearer thanks to the detailed explanation written with references to these figures. This will be understood, and therefore the evaluation should also take these forms and detailed explanations into account. 15 It needs to be done by taking precautions. Figures that will help understand the invention. Figure 1 shows the location of the fault during indoor high-speed tests related to the invention. It is a perspective view of a system to enable its detection. Figure 2 shows the detection of the malfunction during indoor high-speed tests, which is the subject of the invention, and 20 In a method for determining its location, images detected on the tire and It is a schematic representation of the mode of perception. Figure 3 shows the detection of the malfunction during indoor high-speed tests, which is the subject of the invention. One method for determining location is a workflow diagram. Figure 4 shows the detection of the malfunction during indoor high-speed tests, which is the subject of the invention, and 25 In order to determine its location, the change in the tire defect over time and It is the creation of a temporal data sequence. Explanation of Part References 100 Fault and Location Detection Systems and Methods 1 Display Unit 2 Light Sources 3 Tires 4 Drums Body 5 6 Load Cells 7 Processor Units 8 Barcodes Box Sequence 11 Search Trajectory 10 Detailed Description of the Invention This detailed explanation describes the tire in indoor high-speed tests that are the subject of the invention. Method of detecting faults and location (100), preferred configurations, only This is explained to facilitate a better understanding of the subject. Figure 1 shows a tire failure and 15 during indoor high-speed tests related to the invention. The perspective view of the location detection system (100) is given. Accordingly, the fault and The position detection system (100) includes at least one imaging unit (1). The mentioned The imaging unit (1) is essentially a camera and in the indoor high-speed test the tire (3) It is configured to allow viewing. The viewing unit (1) is at least It is associated with a light source (2) to work in conjunction with it. The light source (2) is essentially a 20 It is a strobe system that produces very short flashes of high-intensity light at specific intervals. by generating images that make moving objects appear to slow down or stop. It is an optical technique used for observation and analysis. In a possible configuration of the invention. Light source (2) is provided in a way that there will be 3. At least in the indoor high speed test A tire (3) is tested at predetermined speeds by means of a drum (4). 25 The drum (4) is driven by means of a drive element provided inside at least one housing (5). Variable speed allows the application of variable force via at least one load cell (6). It also provides the variable speed and variable forces applied on the tire (3). 6 with variable slip and variable camber angles in accordance with the test method This ensures that the tire (3) can be used under variable speed and variable load. Testing is ensured. At least the images taken from the imaging element (1) It is necessary to process the deep learning-based system in a processor unit (7). This is the case. Certain information is provided on the tire (3) for the purpose of traceability 5 It contains a barcode (8) which is an information label approved by the FIA. The barcode (8) is essentially a predetermined area between the shoulder and heel of the tire (3). It is connected at this location. In the fault and location detection system, the location of the (100) barcode (8) is accepted as the reference point. It is determined by trajectory passing through the center of the barcode (8) location. It includes at least one search trajectory (11) that is concentric with the tire and has a circle. The aforementioned search trajectory (11) is on a radial scale along the tire radius. The fact that multiple search trajectories (11) are provided by shifting the process step It includes. Detection based on differentiable binary equalization provided on the tire during speed testing 15 At least one barcode (8) location is detected via the model. Barcode (8) by taking its position as a reference (considering it as zero degrees) and it is derivable The bubble representing the final state of the fault is identified using the dilemma method. Sampling over time by setting the position of the bubble as the starting point. A sequence of boxes (10) is obtained by the method. The sequence of boxes (10) is 20 by processing using the convolutional neural network (CNN) method spatial feature array generation, spatial feature array liquid neural network by processing networks (liquid time constant (LTC) / Liquid Neural Network (LNN)) Obtaining the time-dependent feature array, the aforementioned time-dependent feature array 25 The process of obtaining the classification result includes the steps involved. Liquid neural networks (LTC / LNN) detecting only the transition between the robust, early formation and large bubble phases. In doing so, the multi-headed attention block (MHA) identifies which window this transition begins with. is doing. In the fault and location detection system, 30 of the temporally sampled box sequences (100) a sequence of spatial boxes formed by shifting them to a predetermined scale This ensures that the spatial array of boxes is concentric with the center of the rubber band. The definition is provided in the search trajectory (11). The spatial array of boxes, 7 (11) centers in the mentioned fault trajectory are predetermined between them positioned such that the distance is relative to the circumference of the tire (3) The window was provided in the aforementioned search trajectory (11) to the extent possible. Identifying all spatial boxes that have been numbered and personalized as numbers. It has a timestamp during the process and, thanks to window numbers, 5 It is ensured that distance information is included, along with the window number where the fault was detected. spatial boxes with a predetermined sliding distance around the tire circumference The location of the fault is determined by multiplying the angle of change found by the ratio, depending on the barcode position. This allows for the detection of the fault on the tire and the determination of the fault's location. It is ensured that it is located by angular orientation. Tire failure detection and failure time 10 By eliminating dependence on human operators in the creation of the stamp, the correct The results are reported consistently.

Claims

8 REQUESTS 1. Detecting and locating the fault during high-speed indoor tests. It is a fault and location detection system and method (100) and its feature is; tire (3) provided on, At least one barcode (8) location 5 via the derivable dilemma-based detection model detection, The detected barcode (8) is based on time with the position being referenced as zero degrees. a sequence of boxes provided by sampling method (10), By processing the mentioned array of boxes (10) using the convolutional neural network method (CNN) Obtaining the spatial feature array sequence, 10 Spatial feature array generation using liquid neural networks (LTC / LNN) method Obtaining a time-dependent feature array by processing, In the multi-headed attention block (MHA) of the aforementioned time-dependent attribute array sequence obtaining the classification result through processing It includes the steps of the process. 15 2. A method step that conforms to Claim 1 and has the characteristic of being a temporally sampled box. spatial formed by shifting the sequence (10) at a predetermined scale This involves the step of obtaining the array of boxes / kernels / windows.

3. A method step that conforms to claim 1 and whose characteristic is that the spatial array of boxes is rubberized. Concentric identification with the center has been achieved in at least one search trajectory (11) 20 It involves being.

4. A method step in accordance with claim 1, and its feature is the aforementioned search trajectory (11) Multiple failures due to radial scale shifting along the radius of the tire (3). The fact that its trajectory (11) has been ensured includes the process step.

5. A method step that conforms to claim 4 and has the feature of; the spatial array of boxes (10) 25 between the centers of the mentioned fault trajectory (11) a predetermined shifted such that the distance is such that this distance is possible relative to the circumference of the tire (3). In the number given, it includes the step of ensuring that the search trajectory (11) is achieved. 9 6. This is a procedural step in accordance with Request 5, and its characteristic feature is the window where the fault is detected. a predetermined scroll of the spatial array of boxes (10) with its number multiplied by the angle of change of the distance relative to the circumference of the tire (3) The process involves identifying the location of the fault based on the barcode position.

7. A 5 containing a processor unit (7) that implements one of the methods in Requests 1-6. It is a system.

8. Detection of the fault during indoor high-speed tests in accordance with claim 7 and It is a system for determining its location and its feature is that it has at least one light source (2). At least one associated imaging unit (1), from the said imaging unit (1) at least one processing unit (7) in which the acquired images are processed, the mentioned processor 10 At least one AI-assisted decision unit in which tire defects were detected (7) It involves a mechanism.

Images