Abstract: This paper systematically introduces the measurement principle of the suspend gap sensor for Maglev trains. Based on the analysis of the coil configuration, solutions are proposed for the core technical problems of the sensor. Test results show that these solutions effectively solve the difficulties in gap measurement. Keywords: Maglev train, suspend gap, eddy current testing, linearization 1 Overview The suspend gap sensor for maglev trains is an eddy current sensor. It utilizes electromagnetic induction to convert displacement into a change in the coil's self-inductance L or mutual inductance M, which is then converted into a change in voltage or current output by the measuring circuit, thus realizing the conversion of displacement into electrical quantity. This sensor has a simple structure, no moving electrical contacts, high sensitivity and resolution, strong output signal, and good linearity and repeatability. 2. Measurement Principle Since the suspension gap of the EMS maglev train is only about 10mm during operation, and the stator rail itself is a metal conductor, eddy current sensors are the best choice for measuring this small gap. The metal conductor of the eddy current sensor can be considered as a short-circuit coil, which is magnetically connected to a high-frequency energized flat coil. Based on the transformer principle, the high-frequency conductive coil is considered as the primary side of the transformer, and the eddy current loop in the metal conductor is considered as the secondary side. The equivalent circuit of the eddy current sensor can then be drawn as shown in Figure 1. [align=center] Figure 1 Equivalent circuit of eddy current sensor[/align] Where L1 and R1 on the primary side represent the inductance and resistance of the excitation coil, L2 and R2 on the secondary side represent the inductance and resistance of the measured conductor, and R and X represent the resistance and reactance of the load. M is the mutual inductance between the coil and the measured conductor. According to Kirchhoff's laws, the loop equations for the primary and secondary sides are as follows: As can be seen from equation (2), the change in mutual inductance M will cause a change in the impedance of the secondary side refracted to the primary side. From the perspective of the primary coil end, it can be regarded as a change in the impedance of the primary coil. The mutual inductance M is determined by the position between the excitation coil and the conductor being measured, the shape and material of the object being measured. If a constant frequency and constant amplitude signal is used to excite the coil, then the change in position between the coil and the conductor being measured can be measured by detecting the change in the output voltage of the coil. 3 Coil Structure Analysis In the high-speed maglev train system, due to the use of a long stator track, the side of the track facing the sensor is a toothed structure. In the gap measurement process, the effective gap is required to be the distance between the sensor and the tooth surface of the stator track. When the sensor's induction coil corresponds to the slotted structure, incorrect gap information will be generated. When the tooth and slot components corresponding to the coil change, the equivalent inductance of the coil changes, and the final output of the sensor also changes. This is the so-called toothed effect. In order to solve the insensitivity of the coil to this toothed structure, a rectangular induction coil is wound so that its width is equal to one toothed period. As shown in Figure 2(a), during the movement of the coil relative to the stator track, the corresponding cogging component of the coil always consists of one tooth and one slot. Since the magnetic field is strongest at the center of the planar coil and weaker at the edges, we consider adding a compensation coil. This involves adding current-carrying wires to the coil edges to enhance the magnetic induction intensity at the coil edges. See Figure 2(b). [align=center] Figure 2 Comparison of Induction Coils[/align] 4 Cogging Effect Test The test started with a gap of 0mm. At different gaps (every 1mm), the sensor was moved horizontally relative to the guide rail for one cogging cycle, and the output voltage value was sampled (every 6mm was taken) until the gap reached 20mm. We compared the test results before and after adding the compensation coil, as shown in the following figure: [align=center] Figure 3 Test Curves of Two Sets of Coils[/align] As shown in Figure 3, the vertical axis represents the output value of the gap voltage in volts (V). The horizontal axis represents the gap value between the sensor and the stator track in millimeters (mm). The voltage values ​​measured every 1 mm from 0 to 20 mm were correlated with the gap, and the curves of gap versus voltage under different gaps were obtained as shown in the figure above. As can be seen from Figure 3(a), without a compensation coil, the cogging effect of the sensor is relatively large, with a maximum change of nearly 5V. After adding a compensation coil, the sensor's compensation effect for the cogging effect is significantly better than without the compensation coil, as shown in Figure 3(b). 5. Linearization and Output Data Calibration The above test results only reflect the correspondence between the output voltage and the gap. Related processing of the digital signal after A/D conversion is required. According to the protocol, the gap signal from 0 to 20 mm is converted into a digital quantity from 0 to 200. The input signal is divided into several segments, and within a certain local range of the input, the relationship between input and output can be considered approximately linear. When linearizing the data after A/D conversion according to the above formula, we divide it into 10 segments, resulting in a broken line formed by connecting the beginning and end of 10 segments. The independent variable x ranges from 0 to 200, and k is a constant. After plotting the function curve, it is compared with the function y=x (where x ranges from 0 to 200). Here, we define the slope of each segment of the linearized broken line as k1, k2, k3, ..., k10. First, we shift it vertically so that when x=0, we obtain a new function, which is a broken line segment originating from the origin. Then, we divide it into 10 segments and map them to y=x. The final output is: After the above calculation, we take gaps of 0mm, 0.5mm, 1mm, 2mm, 3mm, 4mm, 5mm, 10mm, 10.5mm, 11mm, 11.5mm and 12mm respectively and move the sensor horizontally relative to the stator track by one tooth cycle. The final output result is plotted as shown in Figure 7 below: [align=center] Figure 7 Final Test Curve[/align] As analyzed in the figure above, the linearity and compensation for tooth cogging effect have been controlled within the required range, achieving the expected accuracy. 6 Conclusion The suspension gap sensor, as the core detection part of the maglev train suspension control system, must overcome the negative impact of the cogging effect and the shortcomings of the eddy current sensor itself, and further improve the accuracy of the measurement output. The innovation of the suspension gap sensor prototype we designed and manufactured is that the eddy current coil adopts a new structure, namely a folded and compensated coil. It effectively overcomes the negative impact of the cogging effect and basically meets the accuracy requirements. In summary, our prototype has solved some key problems in sensor design and laid a good foundation for practical application. References [1]. Huang Xianwu, Zheng Xiaoxia. Sensor Principles and Applications [M]. Chengdu: University of Electronic Science and Technology Press, 1999 [2]. Ren Jilin, Wu Liping, Li Lin. Eddy Current Detection [M]. Beijing: National Defense Industry Press, 1985 [3]. Li Lu. Design and Research of Gap Sensor for High-Speed ​​Maglev Train [D]. Changsha: National University of Defense Technology, 2002 [4]. Li Dachao. Internet Interconnection Technology of Embedded Systems [Z], 200 Selected Examples of Embedded System Applications [M]. Beijing: Microcomputer Information Magazine, 2002