Issue 5/2004
09/02/04
The Development of Active Vibration Dampers for Super High-Speed Elevators
Kiyoshi Funai, Hiroko Katayama, Jun-Ichi Higaki, Kenji Utsunomiya and Shinji Nakashima
As super high-rise buildings have been realized in recent years, the requirement of high-speed elevators is getting stronger. For the super high-speed elevators, the reduction of car lateral vibration during their operation is essential for excellent ride comfort. This paper describes the development of active vibration dampers, which realize superior ride comfort.
Category: Issue 5/2004
Posted by: Editor
This paper describes the development of active vibration dampers, which realize superior ride comfort. The characteristics and performance of two types of active vibration damper are described; one is the actively controlled roller guide and another is the actively controlled electromagnetic damper installed between car frame and platform.
In the recent years, high-rise buildings are getting more popular than ever. This trend is driving progress to improve the performance and ride comfort of the elevators that provide vertical transport in tall buildings. Since the elevator cage is small and closed, it is strongly required that the noise and vibration within the cage is small as much as possible not to have the passengers feel uncomfortableness and/or uneasiness. The lateral vibration during the car running is the most important factor that affects the ride comfort. To this end, the development of the technology for the reduction of lateral vibrations especially for high-speed elevators that requires high quality is getting more and more inevitable.
Guide rail deformation and disturbance of air pressure around the car mainly cause the lateral vibrations of the running car. When the height of buildings gets larger and the speed of elevators get faster, the vibrations become much larger due to a large exciting force caused by the above reasons. In order to reduce those vibrations and attain an excellent ride comfort, we introduced oil dampers and other passive devices to the guide shoe or located between the elevator car and the frame. Additionally, the greatest care is taken to increase the precision of guide rails during machining and to improve alignment accuracy during installation to minimize the perturbing force. These passive measures, however, are insufficient to get enough performance to reduce the vibrations for high-speed elevators of high-rise buildings.
To solve this problem, several active vibration dampers have been studied. For example, the method using actuators for roller guide (Utsunomiya et al. 2002), the method of using actuators installed between the car frame and the platform (Kato and Higaki 2002) and active mass damper system (Teshima et al. 2002). There seems to be, however, few reports describing the performance and characteristics to evaluate which method is most suitable for the actual elevators (Funai et al. 2002).
In this paper, we would like to compare the systematic characteristics of the typical two active dampers that we have developed: the method of using actuators for roller guides and the method of using actuators between the car frame and the platform.
Vibrations of the cage
Figure 1 shows a schematic diagram of the structure of elevator cage. Roller guides support the cage with springs to avoid the transmission of the perturbing force caused by the deformation or misalignment of guide rails. The platform rubbers are attached between the cage and the frame to insulate the vibrations.
The elevator cage is regarded as a multi-body vibration system which consists of mass elements and elastic and damping elements. Figure 2 shows the vibration modal shapes (horizontal direction only) of a typical high-speed elevator. There are in-phase modes of car and cage between 1 and 3 Hz, and out-of-phase modes between 4 and 6 Hz.
Active roller guide
Figure 3 shows a photo of the active roller guide we have developed. This device controls the pressing force of the rollers against the rails using actuators, and then reduces the perturbing force caused by the deformation of guide rails. An actuator for each roller generates pressing force against the guide rails. There are three accelerometers attached near the roller guides to detect horizontal vibrations. We installed two sets of active roller guide under the cage, because the lower position is most effective for the reduction of the vibration of the cage floor. They control the force of three degrees of freedom in a horizontal plane (back and forth, side to side and rotation) of the cage.
Since the active roller guide can be applied easily just by exchanging the normal roller guide, it has merit that it can be applied not only to newly installed elevators but also to modernization of old elevators.
Skyhook damper control is adopted for the control rule of the active roller. The skyhook damper control virtually realizes the property that the cage is supported from an absolutely fixed point in the sky. Figure 4 shows a block diagram of the active vibration controller.
Active vibration damper under the cage
Figure 5 shows the schematic diagram of the active vibration damper installed under the cage and a photo of the electromagnet used for an actuator. There are two sets of the electromagnetic actuators, which consist of coils attached under the cage and yokes attached on the frame. Each set of electromagnet is arranged along back and forth axis and side to side axis. They reduce the horizontal vibration of the cage by controlling the electromagnetic force.
Since this system is installed under the cage, it has merit that a large size device can be arranged to attain high power driving force.
For the controller of the actuator, skyhook damper control and skyhook spring control is adopted for this system. When the external perturbing force is very large or the mechanical damping of the system is large, it is sometimes hard to suppress the vibrations enough by only the skyhook damper control. In such a system, skyhook spring control is very effective to suppress the vibrations because the skyhook spring control is equivalent to the system that has a large stiffness by feedbacking the displacement.
Figure 6 shows a block diagram of the controller. Since the actuator of this system utilizes the electromagnetic attractive force, the control is difficult because of the nonlinearity of the electromagnetic attractive force that depends on the distance between the coil and the yoke. In this system, we have developed the linearization filter by compensating the electric current by measuring the distance between the coil and the yoke using an eddy current gap sensor.
Comparison of control characteristics
In order to compare the characteristics of the both active dampers, we used a simple two-degree-of-freedom model to express the horizontal vibration of the elevator cage and the frame shown in Figure 7. In this model, m1, m2, k1, k2, c1, c2, x0, x1, x2 denotes mass of frame, mass of cage, spring stiffness of roller guide, spring stiffness of platform rubber, damping coefficient of roller guide, damping coefficient of platform rubber, displacement of guide rail, frame, and cage, respectively.
Figure 8 and 9 show the vibration suppression characteristics of the active roller guide and active damper under the cage using the above model. In these figures, a broken line shows without-controlled case and a solid line shows with-controlled case. For this evaluation, we only use skyhook damper control to compare the performance of the both systems.
Figure 8 shows the vibration suppression characteristics against the displacement of the guide rails, i.e., closed loop characteristic from the displacement of guide rail to the acceleration of the cage. From this figure, the active roller guide has comparatively excellent suppression performance for the first mode (in-phase mode), while the active damper under the cage has excellent performance for the second mode (out-of-phase mode). The reason of this characteristic is as follows: since the acceleration of the active roller guide is detected at the frame and the controlled force is applied to the frame, the in-phase mode is easier to be suppressed. On the other hand, the out-of-phase mode is not easy to be suppressed because the driving force is just applied on the frame. In the active damper under the cage, since the acceleration of the cage is detected on the cage and the driving force is applied between the cage and the frame, the out-of-phase mode is easier to be suppressed.
Next, we evaluated the performance when the perturbing force caused by the disturbance of the air pressure around the cage is applied to the cage directly. Figure 9 shows the vibration suppression characteristics when the external force is applied to the cage, i.e., closed-loop characteristic from disturbing force of the cage to acceleration of the cage. In the Figure 9, while the second mode is not suppressed enough for the active roller guide, the both modes are suppressed equally for the active damper under the cage.
The difference of the characteristics is caused by the differences of the position of the actuators and the acceleration sensors.
Experimental result
To assure the performance of the both active dampers, we carried out practical trials using actual elevators installed in our experimental tower.
Figure 10 shows the result of active roller guide and Figure 11 shows the result of active damper under the cage. The speed of the elevators is 5 m/s for the active roller guide and 3 m/s for the active damper under the cage. For both experiments, the vibration level is exaggerated by enlarging the deformation at the joint part of the guide rail purposely in order to make the effect of the control clear to understand. The loop gain is determined aiming the vibration level is to be suppressed to a half of the noncontrolled case for the both experiments.
These results show that the both active dampers realize the performance as we expected against the perturbing force caused by the guide rail deformation. The reason that both of the devices have enough suppression performance is that the dominant frequency of the guide rail deformation is less than 1 Hz and the controllers have enough performance of suppression around this frequency as shown in Figure 8.
The frequency of the guide rail deformation depends on the guide rail length and the speed of the elevator car (Utsunomiya et al. 2002).
Aerodynamic simulation
For super high-speed elevators, the disturbance of the air pressure around the cage generated by another cage running next in the same shaft can not be neglected. Then, we evaluated the performance of the vibration suppression against the aerodynamic disturbance quantitatively. Firstly, we made an aerodynamic simulation to estimate the air pressure distribution around the cage while the car is running.
Simulation Model
We made a two-dimensional aerodynamic analysis model of the super high-speed elevators. Using the model, we evaluated the air pressure distribution when the two cars run abreast and pass each other in the same shaft. Figure 12 shows a car model used in the simulation. A very long and sharp streamlined fairing is attached over and beneath the cage. Since the fairing is fixed on the frame, the air pressure applies on the fairing and the cage separately. The car speed of the simulation is 18 m/s.
Analysis Results
Figure 13 shows two of the results of air pressure distribution calculated by the simulation. The figure shows that the vibratory phenomenon of air pressure distribution around the cage affected by another car in the both cases.
Figure 14 shows the force fluctuation applied to the cage due to the airflow disturbance. It would be used as external disturbance force to the cage for simulations of vibration in the next chapter. The frequencies indicated in the figures are the dominant frequencies of the oscillation of the air pressure. The frequency is affected by the airflow speed and the cross-sectional area of the flow.
Figure 14(a) shows the air pressure fluctuation when the car runs abreast. At the beginning of the calculation, there is a rapid force change like a step. We used a stable part (2–4 seconds) for the vibration analysis in the next chapter, because the beginning part is just a transient phenomenon only in the simulation.
Figure 14(b) shows the air pressure fluctuation when the two cars pass each other. When the cars get close, repulsion force acts to each other. After the cars have passed each other, unstable eddy flow is generated behind the cages, and then a large attraction force is generated.
Since the simulation model is two dimensional, it should be noted that the amplitude and the frequency of the vibration of the actual elevators are about a half of the simulation results.
Vibration simulation
For the assurance of the vibration suppression performance of the two active dampers against the perturbing force by air pressure, we have calculated time response using the air pressure simulation result. Figure 15 shows a simulation model of the elevator cage for the lateral vibration. The model is threedimensional 24 degree-of-freedom. Figure 16 shows the simulation result of the vibration of the cage when the two cars run abreast and pass each other. In the simulation, we used the same control gain for both active control systems and the perturbing force is generated only by the airflow disturbance (no guide rail perturbing force).
In Figure 16(a), when the cars run abreast, the active roller guide could not suppress the vibration, rather, increased the amplitude compared to the without-controlled case. This phenomenon occurred because the dominant frequency of the air pressure oscillation (3.7 Hz) is the frequency that the vibration suppression gain is larger than that of without-controlled case in Figure 9. The active damper under the cage suppresses the vibration to almost less than 10 cm/s2 as we expected.
In Figure 16(b), when the cars pass each other, the active damper under the cage (floor active damper) has enough performance of suppressing the vibration caused by the rapid change of the air pressure. On the other hand, the performance of the active roller guide is less than that of the active damper under the cage. The reason is that the frequency of the dominant frequency is 2.2 Hz whose vibration modal shape is close to the out-of-phase mode of cage and frame. The performance of the suppression for the out-of-phase mode of active roller guide is not sufficient as shown in Figure 9.
Thus, we obtained the following result:
• For the both cases of the cars running abreast and passing each other, the active damper under the cage is capable of suppressing the car lateral vibration as much as we expected.
• The active roller guide has inferior performance of suppression compared to the active damper under the cage because the frequency of the air pressure disturbance is higher than that caused by rail guides displacement. Also, since the controlled force of active roller guide does not act on the cage directly but on the frame, the performance is not enough compared to the active damper under the cage whose force acts on the cage directly.
Conclusion
The performance of the vibration suppression of two types of active vibration dampers is evaluated against the guide rail displacement and the air pressure disturbance. From the evaluation, we obtained the following results:
• Both of the active dampers have sufficient performance of vibration suppression against the guide rail displacement caused by rail deformation or misalignment.
• For the air pressure disturbance, the active damper under the cage has excellent performance compared to active roller guide. Because the dominant frequency of the airflow disturbance is comparably higher and close to the out-of-phase mode of the cage and frame vibration.
• For the air pressure disturbance, the active damper under the cage has excellent performance compared to active roller guide. Because the dominant frequency of the airflow disturbance is comparably higher and close to the out-of-phase mode of the cage and frame vibration.
From the above results, we have concluded that the active roller guide is suitable for middle and high speed elevators, which are mainly affected by guide rail deformation much more than air pressure disturbance. On the other hand, the active damper under the cage is suitable for the super high-speed elevators which are extremely affected by air pressure disturbance.
References
Funai, K., Kuraoka, H., Higaki, J., Utsunomiya, K. (2002). The development of active vibration damper for elevators. Elevator, Escalator and Amusement Rides Conference, January 2002, JSME, pp. 13–16 (in Japanese).
Kato, S., Higaki, J. (2002). The world’s fastest elevators. Mitsubishi Electric ADVANCE, September 2002, pp. 19–21.
Teshima, N., Kamimura, K., Kohara, H., Nagai, M. (1996). Vibration control of a ultra high speed elevator by active mass damper. The fifth transportation and logistics conference, December 1996, JSME, pp. 97–100 (in Japanese).
Utsunomiya, K., Okamoto, K., Yumura, T., Funai, K. and Kuraoka, H. (2002). Active roller guide system for high-speed elevators. Elevator World, April 2002, pp. 86–93.
Biographical details
Kiyoshi Funai joined Mitsubishi Electric Corporation in 1979. He is presently manager of Elevator Traction Machine Development Section, Inazawa Works, Japan.
Hiroko Katayama joined Mitsubishi Electric Corporation in 2002. She is presently a mechanical engineer in Elevator Traction Machine Development Section, Inazawa Works, Japan.
Jun-Ichi Higaki joined Mitsubishi Electric Corporation in 1990. He is presently a mechanical engineer in Mechanical Dynamics Section, Advanced Technology R&D Center, Japan.
Kenji Utsunomiya joined Mitsubishi Electric Corporation in 1997. He is presently a mechanical engineer in Mechatronics Department, Advanced Technology R&D Center, Japan.
Shinji Nakashima joined Mitsubishi Electric Corporation in 1988. He is presently manager of Compound Structure Technology Group, Manufacturing Engineering Center, Japan.
First published by IAEE – The International Association of Elevator Engineers in Elevator Technology 14, Proceedings of Elevcon 2004, the 14th International Congress on Vertical Transportation Technologies, held 27–29 April 2004 in Istanbul, Turkey.
*) Mitsubishi Electric Corporation, Japan
5/2004
