Abstract: Currently, domestic rail wear detection technology is mainly limited to three methods: manual inspection, which is inefficient and has large measurement errors; contact electronic equipment measurement, which results in relatively high sensor wear; and non-contact optical measurement, which has high accuracy but limited applicability. To address these problems, this paper proposes a system design for rail wear detection based on virtual instruments. This system mainly utilizes the non-contact detection principle, combining an eddy current displacement sensor with NI's LabVIEW and PXI devices to achieve online real-time detection, processing, and analysis functions. It features low cost, high accuracy, and simple operation. Verification experimental results demonstrate that this system has superior performance compared to other methods. Keywords: rail; wear detection; eddy current sensor; virtual instrument[b][align=center]Design of Rail Wear Detection System Based on Lab VIEW Wang Junfeng Song Wenai[/align][/b] Abstract: The technology of rail wear detection currently in china is mainly developed in three methods: manual detection with low efficiency and many errors, contact electronic detection with high relative wear of sensor and optical detection with high accuracy and low application. So it is very important and necessary to develop a set of detection system with high accuracy and application. This article proposes one kind of design for rail wear detection system with non-contact eddy current sensor and PXI based on Lab VIEW. The system can realize online real-time detection, processing, analysis functions, has advantage in low cost, high accuracy, simple operation and so on. The confirmation experiment results have indicated this system has more superior performance than other methods. Key words: Rail; Wear Detection; Eddy Sensor; LabVIEW 1 Introduction At present, the railway locomotive department in China mainly uses the outdated detection method of manual inspection to detect rail wear. This measurement method is inefficient, unreliable, labor-intensive, and environmentally harsh. In addition, some research institutions have developed new detection technologies, such as electronic equipment detection and non-contact optical inspection. Mechanical equipment requires manual inspection, which has the disadvantages of large workload, harsh working environment, low efficiency, contact between mechanical device and rail, resulting in insufficient accuracy and integrity of the obtained rail shape and gauge, long-term contact measurement causes unnecessary wear on the rail, and the measurement process is difficult due to the need to pass through some accessory connection points. It is also easily affected by the environment. Although electronic equipment has been greatly improved compared to mechanical equipment, online automatic detection has not yet been realized, which increases the railway maintenance cycle. Non-contact optical system detection is better than the above two, with high accuracy, fast response, safety and reliability, stable operation, and avoidance of contact with moving objects, thus avoiding unnecessary wear[1][2][3]. However, the system cost is relatively high and its practicality is not strong. To effectively address the aforementioned issues, this paper proposes a LabVIEW-based system design for rail wear detection. This system primarily utilizes non-contact detection principles, enabling online real-time detection, processing, and analysis. Experimental results demonstrate that this system exhibits superior performance compared to other methods. 2. Measurement Principle As shown in Figure 1: The data measured by the displacement sensor is transmitted to an amplifier and filter preamplifier, converting it into a standard 0-5V voltage signal. The data acquisition card acquires the voltage signal and sends it to the computer. The computer processes the signal acquired by the acquisition card to obtain the wear amount, which is displayed in real time. Simultaneously, the computer can control a motor to drive the sensor to rapidly scan the top surface of the rail, ensuring measurement of the entire top surface. The computer also performs data storage and management. The mechanical hardware measurement principle is shown in Figure 2: The positioning wheel 1 and support wheel 3 are reliably contacted with the rail surface via a compression spring 5. On a standard rail, the reading of the displacement sensor 4 is constant. When wear occurs on the top surface of the rail, the displacement sensor 4 and the top surface of the rail experience relative displacement, the magnitude of which reflects the wear amount on the top surface of the rail. 3 System Composition 3.1 Hardware Composition [4] The displacement sensor adopts the DO-2 non-contact eddy current sensor of Jiangyin Zhongtai Technology, and is equipped with an amplification filter preamplifier, powered by a DC regulated power supply of ±12V. The data acquisition card adopts the PXI6070E multi-function acquisition card of NI, with an input range of ±5V to ±10V, an accuracy of 12 bits, and a maximum sampling rate of 1.25MS per second. The output is -10V to +10V, with an accuracy of 12 bits and a maximum sampling rate of 1MS per second. The PXI device adopts the PXI1002 chassis with a PXI8185 controller. [align=center] Figure 1 System Hardware Composition 1. Positioning wheel 2. Rail 3. Support wheel 4. Displacement sensor 5. Compression spring Figure 2 Schematic diagram of measuring machinery[/align] 3.2 Software Composition This system is developed using LabVIEW of NI, which can not only easily realize complex data processing functions, but also has a user interface with beautiful, interactive and personalized features; its modular design greatly facilitates the modification and maintenance of the program. 4 System Flowchart As shown in Figure 3. Parameter settings include the measurement personnel, date and time, and road section area. After setting, data collection begins. The collected data should be between 0-5V. Data outside this range should be discarded. During collection, a curve fitting method is used to convert the voltage value into a physical quantity, i.e., wear amount, and the wear amount is displayed in real time. After collection, the collected data is processed, stored, and a report of the measurement results can be displayed and printed. The inspection, wear condition, and service life of this section of rail can be queried. The process ends when no further measurements are taken. [align=center] Figure 3 System Flowchart[/align] 5 System Software Implementation The main system interface is shown in Figure 4. [align=center] Figure 4 Main System Interface 5 Curve Fitting Diagram[/align] 5.1 Sensor Static Curve Fitting Using LabVIEW's built-in curve fitting function, different fitting methods are selected to fit the data to achieve the best fit. Initially, the sensor needs to be calibrated. Static data is collected, and displacement-voltage is fitted. Multiple measurements and multiple fittings are performed to obtain the best-fit curve. As shown in Figure 5, the measurement area of ​​the eddy current sensor is more than three times the probe diameter. Therefore, when the probe moves to the edge of the rail, the original calibration curve is no longer valid. The characteristic curve at the edge needs to be recalibrated. After multiple measurement experiments and using different fitting coefficients, a suitable curve was not found. This means that when the probe moves beyond the edge of the rail, a common characteristic curve is no longer found. Therefore, effective measurements are limited to within the rail edge. 5.2 Data Acquisition and Reading During data acquisition, the actual wear amount is displayed in real-time in both graphical and numerical form. An alarm is displayed if the wear amount exceeds the predetermined maximum wear value. Reading data also displays the wear amount in both graphical and numerical form. The curve can depict different distributions of rail wear, as shown in Figure 6. 5.3 Outlier Removal and Filtering In actual measurement, due to external interference, outliers inevitably appear in the measured data. Therefore, after acquisition, outlier removal and filtering are performed on the acquired measurement data. Outlier removal can be programmable. Filtering can use LabVIEW's built-in filtering functions to achieve the best filtering effect. [align=center]Figure 6 Data Acquisition Module Figure 7 History Recording and Printing Module Front Panel[/align] 5.4 Data Saving To prevent the collected data from being read in an unordered and incomplete manner, resulting in buffer overflow and loss of useful data, the data is automatically saved. Data can be saved in text or report format for easy future data retrieval and historical data querying. 5.5 History Recording Query Using the Database toolkit, historical data can be queried programmatically, by keywords such as time, wear level, and measurement interval. After the query is complete, you can choose to display or print. As shown in Figure 7. 5.6 Report Printing The report printing function can be implemented in the history recording query module. Query results are printed according to the query method. 5.7 Help This mainly provides explanations of system operation and possible problems. 6 System Functional Features The main dynamic functions of this system are: (1) Automatic detection of rail dimensions (2) Data storage (3) Automatic judgment, over-limit alarm, and manual assistance in judgment. The main management and processing functions of this system are: (1) Report display of detection data (2) Automatic establishment of a database of detection results (3) Providing functions such as searching, querying, and statistical analysis of detection result data (4) Report and curve printing functions. 7 Conclusion After multiple verification tests, the experimental results are highly reliable and have basically achieved real-time online detection. Data acquisition and data processing can be carried out simultaneously. The system has high accuracy and is easy to operate. References [1] Meng Jia, Gao Xiaorong. Current status and development of rail wear detection technology [J]. Railway Technical Supervision, 2005, (1), 34-36 [2] D.HolfeId and GE Ghafe. How computer technology aids in measuring rail wear. Railway Track & Structures. 1989, (11) [3] Zhuo Hongyu. Hand-push type rail shape dimension image detection system [D]. Chengdu: Southwest Jiaotong University, 2005.1-15 [4] Zhao Shenghui, Liu Ping, Shi Baohua, Chen Tangxian. Computer control experimental system based on virtual instrument [J]. Microcomputer Information, 2006, (19) [5] Yang Leping, Li Haitao, Yang Lei, LabVIEW Program and Application [M]. Beijing: Electronic Industry Press, 2001 [6] National Instrument Corporation, LabVIEW User Manual. 1998