write in front
The article Active suspension in railway vehicles: a literature survey published in Rail. Eng. Science comprehensively summarizes various important implementations of active control solutions in the field of rail vehicles before 2020. This blog is a faithful translation of the original text of the paper. Since the paper is long, it is divided into several blogs for translation and annotation respectively. This is the fourth article in a series of translations. The previous article is "Active Suspension in Rail Vehicles: Literature Review (3)" .
Paper translation
1 Introduction
2 Basic concepts and classification of active suspension
3 tilt train
4 Active secondary suspension
5 Active primary suspension
5.1 Rigid wheelsets and independently rotating wheels
For traditional vehicles with passive suspension systems, even if a passive guidance mechanism is implemented, a compromise must be made between stability and curve passing performance [90]. In contrast, by applying active suspension to control wheelset motion, it is possible to simultaneously provide a flexible solution to ensure stability and good curve-passing performance. It is expected that the application of active primary suspension will produce greater economic benefits than active secondary suspension because of its relationship with wheel-rail contact. Wear between wheel and rail will be significantly reduced and rolling contact fatigue will tend to improve, resulting in cost savings for vehicles and infrastructure [91].
Wheel sets are generally divided into two types based on their mechanical structure: rigid wheelsets and independently rotating wheels. For a rigid wheelset, both wheels on the same axle will have the same rotational speed. Therefore, longitudinal creep can be generated and the ability to guide and self-centering can be achieved. However, longitudinal creep can cause snaking instability and unnecessary wear of wheel rails when passing through curves. In contrast, independently rotating wheels are free to have different rotational speeds, resulting in a loss of longitudinal creep and ultimately loss of self-centering capabilities, increasing the risk of rim contact.
Generally speaking, when designing active primary suspension, for rigid wheelsets, two issues need to be considered: 'stability' and 'steering'; for independent rotating wheels, you also need to consider "guidance".
Due to the essential difference between rigid wheelsets and independently rotating wheels, it is intuitive to split the active primary suspension section into two parts. In Section 5.2, active solutions for rigid wheelsets are presented, while in Section 5.3, solutions for independently rotating wheels are described. A total of six configurations were considered and they are summarized in Table 1.
5.2 Active primary suspension of rigid wheelset
5.2.1 Principle and configuration
(a) Actuated solid-axle wheelset (ASW)
For rigid wheelsets, ASW is the most widely studied configuration, which can directly apply oscillating torque or lateral force to the wheelset to control oscillating and lateral motion, thereby improving curve passing and stability [3]. This principle can be implemented through three conventional mechanical layouts. The oscillating torque can be applied directly by a oscillating actuator mounted between the bogie and the wheelset, as shown in Figure 22a, or in a more practical way using two actuators in the longitudinal direction at the end of the wheelset, as in As shown in Figure 22b. Since the oscillating motion and lateral motion of the wheel set are coupled, applying a lateral actuator (see Figure 22c) is another way to achieve wheel set motion control. However, a study based on a simplified two-axis vehicle model [92] concluded that lateral actuators require larger forces to achieve the same vehicle stability as a oscillating head actuator. In addition, when using this scheme, it can be expected that the ride quality will deteriorate and the installation space of the lateral actuator will become an issue. Among the three schemes, scheme (b) is the most popular, from which some new specific schemes are proposed, as shown in Figure 23.
Figure 22 General mechanical layout of ASW
Figure 23 Examples of guidance concepts: Actively guided bogie with two actuators [93] (left) and actively guided bogie with one actuator [94] (right)
The scheme in the left panel of Figure 23 was introduced by Park et al. [93], with each wheelset having a longitudinal actuator. Mechanical linkages are designed to transmit the opposing forces of execution at the left and right ends of each wheelset. Therefore, only one actuator is installed per wheel set. The scheme in the picture on the right was proposed by Umehara et al. [94]. A set of carefully designed connecting rods couples the motion of the front and rear wheel pairs, so each bogie only needs one actuator to achieve the steering effect. By introducing articulated mechanical linkages to reduce the number of actuators, installation space and actuator system costs can be saved. However, reducing the number of actuators means that each actuator requires higher performance requirements, such as greater maximum force and maximum pressure, but still has the potential to improve the reliability of the entire system. When reducing the number of actuators, installation space and costs can be saved for fault-tolerant actuators with redundant structures. This is an important aspect in the actual design process [95].
In the “yaw relaxation” concept [96], as shown in Figure 24, a spring is connected in series with the longitudinal actuator, connecting the axle to the bogie frame. On straight tracks, stability is ensured by passive springs and highly stiff actuators operating in a low-bandwidth frequency range or passive mode. In the curve, the actuator is in active mode and can drive the wheelset with less force.
Figure 24 Yaw relaxation plan[96]
If the primary spring is connected in parallel with the longitudinal oscillating actuator to ensure vehicle stability, a higher actuating force is required in the curve to offset the effect of the passive spring. However, parallel passive suspension next to the actuator is an effective way to ensure the fault tolerance of active suspension , which is crucial for the implementation of ASW [95]. When designing a primary suspension, it is difficult to completely remove traditional longitudinal stiffness due to the presence of coil springs or rubber springs that need to bear vertical loads. Nonetheless, reducing the conventional longitudinal stiffness helps achieve lower execution forces, as shown in [97, 98].
(b) Secondary yaw control (SYC)
SYC was originally proposed by Diana et al. to improve the stability of the curve driving performance of straight track and tilting trains [99, 100]. Where the passive oscillation damper was originally installed, the oscillation torque generated from the car body to the bogie is generated by two longitudinal electromechanical actuators. This solution is also known as an active oscillating damper (ADD, German: 'aktiver Drehdampfer'). A schematic diagram of this concept is shown in Figure 25.
Figure 25 Secondary shaking control (SYC)
The SYC concept increases the vehicle's critical speed and reduces track shift forces. Since the movement of the wheelset is uncontrolled, the guiding effect is not as effective as ASW, but the improved stability can allow for a lower primary head angle stiffness, resulting in improved curve performance. Although it is reasonable to classify SYC as an active secondary suspension, the goal of SYC is to improve stability and reduce track-shifting forces in curves, not to improve ride quality. Therefore, this control scheme is closer to the properties of active primary suspension.
Similar to the concept of SYC, [101] proposed a new bogie active control to improve stability. In this work, two inertial actuators are applied transversely on the front and rear beams of the bogie frame. Simulation results demonstrate improved stability, and a recent scaled-down roller bench test demonstrated reduced bogie lateral displacement [102].
(c) Actuated yaw force steered bogie (AY-FS)
Based on SYC, Simson proposed a new active suspension AY-FS for heavy-duty locomotives [103–105]. In this concept, a force steering linkage combined with SYC is implemented. It can be seen as a combination of SYC and passive steering linkage, through which the wheelset can be forced into an ideal position, according to the kinematic relationship between the bogie and the body, as shown in Figure 26. This concept can significantly improve the curve-passing performance of high-traction locomotives.
Figure 26 AY-FS[104]
5.2.2 Control strategies for steering and stability
Control strategies vary according to different control objectives and can be divided into two major categories:
(a) control strategies for steering
(b) control strategies for hunting stability
a. Guidance control strategy
The basic goal of implementing active primary suspension is usually to improve curve performance. The wear number/wear index and the equality of track deflection forces between different wheel pairs are often used to evaluate the curve behavior of a vehicle. The different steering principles presented in the literature are summarized as follows: (a.1) Radial control, (a.2) Perfect steering, (a.3) Locomotive steering control, (a.4) Other controls
a.1 Radial control Radial control
The idea of radial control is to guide each wheelset to a radial position in the curve. In other words, the angle of attack of the wheelset should be as small as possible. Based on this idea, some solutions have been proposed to guide the bogie through passive linkage or wheel pair coupling. For example, the design of Talgo is a good example [106]. However, in theory this control concept can only produce perfect curve passing performance when the superelevation is less than zero, which rarely happens in actual operation. Under normal conditions, a moderately small angle of attack is required to generate lateral creep forces that can balance uncompensated lateral forces. Despite its simplicity, this control scheme has been shown to provide significant improvements in curve behavior [107].
a.2 Perfect steering control
Perfect guidance defined by Goodall and Mei [59] states that the longitudinal creep force of a wheel on the same axis should be equal to zero if no traction or braking force is applied . At the same time, each wheel corresponds to an equal lateral creep force . However, in actual operation, it is very difficult to directly measure the creep force. Therefore, some equivalent indicators that can be used for perfect steering conditions are proposed and summarized.
- Perfect control based on wheelset lateral position
Zero longitudinal force (without taking into account traction and braking forces) and equal lateral force means pure rolling at each wheel . To achieve pure rolling, a head-shaking moment can be applied to control the lateral position of the wheelset [108, 109]. Under the assumption that the wheels are tapered [93], the required lateral displacement of the wheelset can be calculated according to equation (7):
Where e is the track half gauge; r0 represents the rolling radius; λ refers to the wheel taper, and R is the curve radius. Under such control, it may be difficult to measure or estimate wheel taper and the lateral displacement of the wheelset. This method is used for two-axle vehicles, where scaling is performed based on vehicle speed and estimated curvature to improve the stability of the PID controller under different operating conditions [110].
- Perfect control based on yaw moment applied from primary
suspension
This control strategy was originally proposed in the yaw relaxation method proposed by Shen and Goodall [96], followed by some follow-up studies by Perez and Shen [108, 109] respectively. The force exerted on the wheelset can be divided into two parts: the force from the wheel-rail contact and the force transmitted from the bogie through the primary suspension. Creep forces on the two wheels pointing in opposite directions produce a head-shaking moment, but under ideal steering conditions this moment should be zero. If the inertial force of the wheelset is ignored, the head-shaking moment generated by the primary suspension should be reduced to zero. In other words, in order to put the wheelset in a pure rolling position, the oscillation motion of the wheelset is set to be free. Therefore, this control is also called yaw relaxation. The measurement of this head-shaking moment can be achieved by calculating the longitudinal force exerted on the axle box. The actuating force can be measured by the actuating system, and the spring force can be measured by measuring the deformation and stiffness characteristics of the spring. Since the stiffness changes at different load levels, measurement errors may be introduced.
- Perfect control based on ideal angle of attack
The equal ideal angle of attack for each wheel pair is another indicator of the equality of the lateral contact forces of the two wheels on the bogie [109]. The ideal angle for each wheelset is determined by track data (curvature and superelevation), lack of superelevation (vehicle speed) and creep coefficient. Feedforward control can be implemented to simplify control design and avoid system instability. However, many inaccurate measurements of these parameters and uncertainties in the primary suspension stiffness may lead to ineffective control. Therefore, the application of feedback control will improve the effectiveness of steering.
- Perfect control based on same position/movement of wheelsets
In this control strategy, neither an ideal angle of attack nor an ideal lateral displacement is required. Instead, the zero difference between the wheelset's angle of attack and lateral displacement is used as the control target. Forces or deformations of the primary suspension in the lateral and longitudinal directions can be measured as an alternative solution to eliminate differences in wheel set motion.
Overall, the four strategies described above are based on the same principles regarding creep forces. Under ideal conditions where all parameters can be accurately measured, all four control strategies provide perfect steering effects. If sensing issues and measurement uncertainties are taken into account, the steering effects of these control strategies may deteriorate and take on different characteristics. Table 2 compares the above control schemes.
The above four control strategies are proposed because it is difficult to measure creep force. As some specific filtering methods have been proposed for estimating creep force [111], the estimated creep force can be directly used as a control target to achieve perfect steering control [98].
a.3 Steering control in locomotives
Although the perfect guidance in (a.2) can be applied to a wide range of railway vehicles, it may not be suitable for heavy-duty locomotives. Due to the significant longitudinal traction effect, heavy-duty trains in curves will generate longitudinal and transverse forces on the coupler , thus producing a shaking moment [103]. However, in the concept of perfect steering, which requires equal longitudinal and transverse creep forces, this moment cannot be balanced . Taking this into account, Simson et al. proposed two locomotive guidance principles. The "Modified perfect locomotive steering" was first proposed as a modification of the perfect steering . It requires equal longitudinal creep forces on each wheel and equal lateral creep forces on the wheel pairs in the same bogie. In other words, it allows for different lateral forces from the front and rear bogies in order to balance the head-shaking moments. However, the equality of longitudinal forces will limit the utilization of adhesion because the vertical load will vary on each wheel in many operations. In order to solve this problem, "ideal locomotive steering" (Ideal locomotive steering) was proposed, which relaxed the requirement of equality of longitudinal creep force, but its direction should be consistent with the traction force. This strategy allows for a more flexible positioning of the wheelset and can minimize creep forces. Furthermore, wheel rim contact can be effectively avoided even in small curve radii, which is said to reduce the risk of derailment [103].
a.4 Other controls
For the SYC scheme, reducing/balancing the track lateral forces on both axes on the same bogie is the main control objective. A reference execution force is obtained from a lookup table derived from a large number of simulations of different operating scenarios when the vehicle passes through a specific curve radius with a specific uncompensated acceleration [100].
Regarding the AY-FS solution, the reference shaking angle of the bogie can be calculated based on the bogie's position in the curve. A PID controller is introduced to achieve the target angle. Control concepts based on sensing longitudinal forces can be found in the literature [104, 105].
b. Control strategies for hunting stability
The self-excited oscillation of the rigid wheelset introduces snaking instability into the vehicle system. In addition to steering, achieving stability is the main goal of active primary suspension. The wheel set's yaw angle, lateral speed and yaw angular speed are three indicators to measure the instability of the wheel set. Three corresponding control strategies are proposed to stabilize the motion of the wheelset.
Active lateral damping and active yaw damping are two similar control strategies. The lateral force exerted by the former is proportional to the oscillation angular velocity of the wheelset, while the oscillation moment introduced by the latter is proportional to the lateral velocity of the wheelset. The stabilizing effect of these two control strategies has been theoretically verified in the literature [3] based on the two-axle vehicle model. Active oscillation damping is superior to active lateral damping because it requires lower forces and produces better ride quality [92]. The active shaking head damping effect was also verified in the test of Pearson et al. [112].
The third stability control strategy is called 'Sky-hook spring' (also known as 'Absolute yaw stiffness'). Inspired by the ineffectiveness of passive head-shaking stiffness [44, 113]. In a passive primary suspension, the head-shaking moment generated by the spring is proportional to the relative rotation angle between the bogie and the wheel pair, but the ideal head-shaking moment should be proportional to the absolute head-shaking angle of the wheel pair . Increasing the stiffness of the springs can enhance the effect of the head-shaking moment and thus improve stability, but the effect will still be affected due to the movement of the bogie. To solve this problem, “canopy springs” were proposed [113]. The head-shaking force acting on the wheelset is proportional to the measured absolute head-shaking angle of the wheelset, and a high-pass filter is used to remove low-frequency signals during curve driving. Table 3 shows the above three control strategies.
In SYC, vehicle stability is achieved through two longitudinal actuators used to mimic secondary yaw dampers (two longitudinal actuators used to mimic secondary yaw dampers). The reference force of the actuator can be calculated according to formula (8):
where vrel is the relative speed between the bogie and the car body, and cv is a gain similar to viscous damping. By introducing the mvax term, the delay of the sensor and the action of the actuator are compensated [100].
The control principles of stability and guidance have been summarized separately, and control of actuator behavior can also be achieved by using a controller like H∞. The efficiency of H∞ controllers has been studied on conventional bogie vehicles (Mousavi et al. [114]) and two-axle vehicles (Qazizadeh et al. [115]), where it was shown that curve performance can be achieved while reducing execution forces improvement.
5.2.3 Examples
Apart from the concept of SYC, there are currently no commercial applications of active primary suspension, but significant new progress has been made since the last review of work in 2007 [4].
SYC has been tested on full-scale roller test rigs and tracks in the early 2000s [99, 100, 116]. Tests confirmed the reduction of wheel set guidance forces and the improvement of vehicle stability. Siemens has applied ADD to electric locomotives (models ES64F4 and ES64U2 [117]) and plans to use this system on the new generation locomotive "Vectron".
In recent years, the ASW concept has received more attention. The following is a description of roller test rigs and field tests conducted in different countries.
Umehara et al. [94] developed a bogie frame based on the scheme shown on the right side of Figure 23. Each bogie has only one electrohydraulic actuator (EHA), and through special circuit and valve design, the risk of any actuator inverse control is eliminated [118]. If reverse guidance occurs, the actuator will produce no force, i.e. the actuator system is safely disabled. Tests were conducted on a straight track at a speed of 10 km/h to verify the safe failure function of the EHA actuator, and the results showed that the wheel set guidance force did not increase under reverse guidance control.
Around 2010, the Korea Railway Research Institute conducted experiments on a 1:5 scale roller test rig based on the mechanical layout shown on the left side of Figure 23 [93]. Perfect steering control of the wheelset based on ideal lateral displacement is implemented, assuming a fully tapered wheel tread. An electromagnetic linear actuator was used in the experiment. Tests showed that improvements were achieved in terms of wheel set lateral displacement and guiding forces.
Following this, a series of studies focused on new steering layouts, ranging from simulation to field testing. The control strategy is based on the radial position of the wheelset in the curve introduced in Section 5.2.2 (a.1). proposed a real-time curve radius estimation method where the only measurement was the relative displacement between two points from the car body and the bogie [97]. Special electromechanical actuators were created that drive the rotation of the motor together with the connecting rod to achieve linear motion of the rod at both ends, as shown in Figure 27 (left), unlike traditional linear actuators. In this way, there is only one actuator unit per side of the bogie. In addition, the primary rubber springs were redesigned to achieve lower longitudinal stiffness. Figure 27 (right) shows a prototype of a steering bogie that was subjected to static experiments on track for the first time to examine the motion of the actuator system [119]. The curvature signal measured in advance was sent to the control system to simulate the real track information, and then the yaw angle and driving force of the wheelset were measured. Considering the maintenance and cost of the head angle measurement system, feedforward control is used to control the displacement of the actuator. Therefore, the measurement of the wheelset head angle is only used to evaluate the wheelset motion. The maximum error between the reference guide angle and the measured value is 8%. In addition, the controller has self-diagnostic capabilities. When an error signal is recognized, the actuator switches to passive bogie mode.
Figure 27 Guide actuator device [97] (left picture) and guide bogie prototype [107] (right picture)
A recent publication [107] presents the results of experiments conducted on commercial lines. The accuracy of curvature estimation and steering angle is satisfactory, with maximum errors of 2.4% and 4.9% respectively. The lateral forces on the wheels were significantly reduced, and tests over a total of 1,000 kilometers showed negligible rim wear.
In China, CRRC demonstrated a prototype of the next-generation metro vehicle “Cetrovo” equipped with an active steering system [120]. Dr. Wang Xu, senior engineer at the R&D Center of CRRC Qingdao Sifang Rolling Stock Co., Ltd., showed us the recent progress. The steering system uses a hydraulic servo actuator . The reference displacement of the actuator is calculated based on the track curvature obtained through a track curvature database and geolocation technology . Preliminary field testing has recently been completed, showing significant improvements in vehicle guidance capabilities and wheel-rail contact noise.
Bombardier has developed a double-decker train called TWINDEXX in which active steering is implemented. However, no further technical information has been released so far and the current development status is unclear.
It is important to note that when approaching the final implementation of active primary suspension, the fault tolerance of active suspension systems should be carefully considered . While beneficial effects have been identified, safety-critical issues must be properly addressed at the same time. Some safe failure designs have been carried out in the above work. In the future, attention should continue to be paid to the fault-tolerant design of execution systems.
(unfinished)
For follow-up content, please continue to read the blog post: