1. Introduction To improve the ride comfort of the Maglev train, control techniques are important. Three control techniques were introduced into the Yamanashi Maglev Test Line vehicle. One method uses SCMresiliently-mounted bogies. The second method uses a full-active suspension system of oil hydraulic actuators, and the third method a semi-active suspension system of adjustable (damping force) damper, both set between the car body and bogie. This paper reports an overview of the application of the semi-active suspension system to the Yamanashi Maglev Test Line vehicle. An improvement of the ride comfort levels was confirmed by vehicle running tests. 2. Semi-active suspension system 2.1 Control law In this system, the control law adopted a sky-hook damper control. The approach is shown in Figure 1. Fig.1 Sky-hook damper As sources of vibrations of the car body, the next two types are considered: 1). Vibrations that the bogie transmits 2). Vibrations that the car body is subjected to aerodynamically If a damper can be fixed to a stationary virtual sidewall as in Figure 1, it is possible to reduce car body vibrations of both cases by this damper. This imaginary damper is called a sky-hook damper. The control based on this is called the sky-hook damper control. 2.2 Features of the semi-active suspension system To transmit the force which arises in the sky-hook damper, there is a method for supplying energy externally such as by hydraulic actuators, and a method using adjustable dampers. The former is a full-active suspension system, and the latter is a semi-active suspension system. In the case of the semi-active suspension system, when the relative motion between the car body and bogie can not generate the force in the sky-hook damper in the direction as commanded by the control law, the generated damping force is set to zero. It might be considered inferior to the full-active suspension
system, which can output the sky-hook damper force regardless of the relative motion between the car body and bogie. However, depending on the type of object to be controlled, the semi-active suspension control can reduce vibrations nearly as effectively as the full-active suspension system. In addition, the semi-active suspension control has the following features of being excellent for rolling stock: (1) It is compact and light in weight because no mechanical power source is necessary. It is also easy to install and perform maintenance. (2) It is highly stable. It is a fail-safe system that simply reverts to conventional damper characteristics in emergency cases when the control system is deactivated. (3) It is lower in cost than the full-active suspension system. (4) The adjustable damper can be replaced by a passive damper. Remodeling of the car body and bogie upon the introduction of the control system to the vehicle is unnecessary. 3. Performance of the control system predicted by simulation In the case of the Maglev system, magnetic springs work between the bogies and the ground. Therefore, Maglev demonstrates movements unlike those of the Shinkansen. The performance of the control system was estimated by simulation. The vehicle model is a three-car train set. Irregularities in the position of coils on the sidewalls of the guideway of the Yamanashi test line were input as a disturbance, and the controlled object was set to be a damper of an articulated bogie. As a result of the calculations, the following facts were proven. (1) There is improvement in the 1-6 Hz frequency domain. (2) The decrease in the riding comfort level was about 3.0 db. If there is 3 db of improvement, the effect can be clearly noticed. Therefore, improvement in riding comfort was predicted for the semi-active suspension system of the Maglev vehicle. 4. Issues of development unique to the Maglev system 4.1 Countermeasures for the effects of magnetic fields on the adjustable dampers Countermeasures for the effects of magnetic fields are indispensable for each equipment used in the Maglev system. The semi-active suspension system is composed of adjustable dampers, accelerometers of the car body and a semi-active controller. These countermeasures are especially important for the adjustable dampers, since they are installed in the location of particularly strong magnetic fields in the bogie. There are currently two types of adjustable dampers that can be used: the "high-speed solenoid valve system" used in Shinkansen and the "proportion relief valve system" developed as a low-cost type. In the case of the high-speed solenoid valve system, it is possible to choose the damping coefficient from six different settings, by changing the orifice using three high-speed solenoid valves. This system can vary the damping force very quickly, and has quick response. Though the proportion relief valve system is lower in cost, the response is slower than in the high-speed solenoid valve system. In the case of the Maglev system, the high-speed solenoid valve system which has quicker response is more advantageous for car body vibrations, since these vibrations include a large proportion of higher harmonics. However, in case of the high-speed solenoid valve system, detection of the piston speed is necessary. In the existing system, a stroke sensor is used in order to obtain the piston speed. This stroke sensor uses a magnetic sensor head to read a magnetic scale that is marked on a rod. Therefore, it is very unlikely that this stroke sensor can function properly in magnetic fields. On the other hand, in the proportion relief valve system, the stroke sensor becomes unnecessary as a result of the design of the damper. The high-speed solenoid valve, stroke sensor, proportion electromagnetic relief valve and unload valve were tested in magnetic fields to confirm that they function properly, in order to decide on the system of adjustable dampers for the Maglev vehicle. It was concluded from test results that the proportion relief valve system composed of proportion electromagnetic relief valves and unload valves could be used for the Maglev vehicle. 4.2 Countermeasures for the effects of magnetic fields on the accelerometers In the sky-hook control, the car body speed is indispensable data, obtained by integrating the output
of accelerometers. Currently there are two types of accelerometers used for semi-active suspension systems: servo mechanical accelerometers and semiconductor strain-gauge system accelerometers. Since the servo type is highly likely to be influenced by magnetic fields as would be indicated by offsets of the output, the semiconductor strain-gauge type which is unlikely to be influenced by magnetic fields was adopted for the Maglev vehicle. Moreover, the accelerometers are installed inside the car body, to ensure that the effects of the magnetic fields are minimized. Because the acceleration sensor output is an important signal in the control system, accelerometers were installed inside the Maglev vehicle and actually tested to ensure that they function properly. The results showed that the accelerometers posed no problems and that they were suitable for use in the control system. 5. Semi-active control system for the Yamanashi test line vehicle 5.1 Adjustable damper Considering the factors discussed in the previous section, the adjustable damper for the Yamanashi test line vehicle was constructed. The appearance of the adjustable damper for lateral vibrations used in the Maglev vehicle is shown in Figure 2. Fig.2 Appearance of the adjustable damper for lateral vibrations The adjustable damper for lateral vibrations will not interfere with the relative motion between the car body and bogie, even if it were replaced with a passive damper. However, the adjustable damper for vertical vibrations interferes with the relative motion between the car body and bogie when it is replaced with a passive damper. Therefore, the bogie had to be remodeled. With the exception of some of the abrasion-resistant parts, the components were constructed of aluminum and stainless steel to ensure that the damper is non-magnetic (diamagnetic). In order to observe characteristics of the adjustable damper for the Yamanashi test line vehicle, this damper was installed in a horizontal servo testing equipment to measure the damping force. The relationship between the piston speed and the damping force characteristics were observed when the command voltage was input to this damper. The results confirmed the following: (1) The damping force characteristics of this damper are nearly equivalent to the design. (2) When the power is cut, the characteristics of this damper are almost equal to that of a passive damper. 5.2 Accelerometer Semiconductor strain-gauge system accelerometers were selected to be used in the Maglev vehicle, as was mentioned above. The appearance of the accelerometer used by the Maglev vehicle is shown in Figure 3.
Fig.3 Appearance of the accelerometer In the semi-active suspension system, the car body speed is obtained by accelerometers. A command force corresponding to the car body speed is output. Therefore, if the accelerometer malfunctions such that the condition of zero car body acceleration continues, the command force also becomes zero. Consequently, an abnormally low damping force continues to be output, leading to abnormal and dangerous motions of the vehicle. Therefore, the semi-active suspension system must have fault diagnosis capabilities built into the accelerometer itself. By devising an appropriate configuration for the sensor, the fault diagnosis was built into the accelerometer of this system. 5.3 Semi-active controllers and the composition of the semi-active suspension system The composition of the semi-active suspension system for the Yamanashi test line vehicle is shown in Figure 4. Fig.4 The composition of the semi-active suspension system
The controlled object is the body of the middle car in a three-car train set. Because the Yamanashi test line vehicle has adopted an articulated bogie arrangement, the motion of the adjacent car body may interfere in the control of the middle car body. Therefore, all the passive dampers of the bogies attached to the middle car body were replaced by adjustable dampers. Though one controller fundamentally controls one bogie, the transmission of signals is possible between two controllers, enabling control of yaw and pitch motions of the car body. The appearance of the semi-active controller is shown in Figure 5. Fig.5 Semi-active controller 6. Tests on the Yamanashi Maglev Test Line The semi-active suspension system was installed in the Yamanashi test line vehicle, and vehicle running tests were carried out on the Yamanashi Maglev Test Line. 6.1 Lateral vibration reductionperformance by the control system The decrease in the interval average of the lateral riding comfort level using the control system was about 2-3 db. When the short-time riding comfort level analysis was carried out, an improvement in the riding comfort level of 2-5 db was confirmed. The following was proven as a result of PSD analysis of the car body lateral vibration acceleration: (1) The PSD is reduced by the control system for frequencies of 3 Hz or less. (2) Although peaks indicating no improvement exist in the 4-6 Hz range, these peaks are considered to be stationary points which originate from the bogie resonance. (3) The human body was able to perceive the difference in riding comfort using the control system. 6.2 Vertical vibration reduction performance by the control system The reduction of vibrations by this control system is perceived by the human body and therefore very effective. When the short-time riding comfort level analysis was carried out, there was an improvement in the riding comfort of about 2 db at the peak. However, there was very little decrease in the interval average of the riding comfort level. As a result of the PSD analysis, the following were proven: (1) The car body vibrations of the first bending mode at 13 Hz is excited by the control system. (2) The car body vibrations of the pitching mode at 4 Hz is not reduced. These characteristics seem to influence the vertical riding comfort level. In order to improve the damping effect, the response speed of the semi-active damper will be improved, and a cooperative control system that processes measurements from devices attached to the car body and the bogies at both the front and rear ends of the car body.
7. Conclusions A semi-active suspension system for the Maglev vehicle was developed which can function properly under strong magnetic fields. Results of tests on the Yamanashi Maglev Test Line showed that there was an improvement in the interval average of the riding comfort level of about 2-3 db for lateral vibrations, but there was little improvement for vertical vibrations. To reduce vertical vibrations, the damper response speed will be improved. Acknowledgment The authors would like to express their sincere gratitude to Mr. M. Nakazato, Mr. J. Arai, and Mr. M. Uchida of KAYABA INDUSTRY CO., LTD., whose time and efforts in helpful guidance were indispensable to the completion of this paper. The authors also thank Mr. Erimitsu Suzuki for correcting the English. This work is financially supported in part by the Ministry of Land, Infrastructure and Transport, of the Japanese government. BIBLIOGRAPHY K. Sasaki, S. Kamoshita, T. Shimomura "Semi-active suspension for rolling stock," RTRI Report (in Japanese), Volume 10, Number 5, pp 25-30, May, 1996. K. Sasaki, "Control technology for riding comfort improvement," RTRI Report (in Japanese), Volume 13, Number 10, pp 1-6, October 1989.