Design, Control, and Comparison of Low-Energy Solenoid Valve Actuators

Abstract

An automotive, fluid-control solenoid valve is composed of an electromagnetic reluctance actuator and a near-constant-force spring. Reluctance actuators are applied as electromagnetic brakes in aerospace applications [1], as valves that perform fast sorting tasks by means of short air-pulses in the manufacturing industry [2], as accurate fluid-control valves in petrochemical processes [3], and in the automotive industry to achieve variable valve timing in camless engines [4]. Common desires are a fast switching and low noise upon impact. Preferably, these objectives are met with minimized energy consumption, especially during constant position operation. In addition, minimizing the impact velocity improves valve lifetime and reduces the audible noise, vibration, and harshness (NVH). This paper considers cylindrical reluctance actuators due to their low cost. However, this complicates the use of laminations to minimize eddy current effects in a cost-effective manner. Proper analysis, design, and optimization of the reluctance actuator can, therefore, only be performed if these dynamic effects in the actuator are accounted for. This paper will focus on incorporating the eddy current effect in the models and their effect on performance, as well as control methods to improve the performance and minimize energy consumption. The performance of a classical reluctance actuator (Fig. 1a) is compared to a PM-biased topology (Fig. 1b) which reduces the energy consumption. Modeling is performed using transient, axisymmetric, nonlinear finite element (FE) simulations, coupled to Matlab-Simulink. Actuator topology and constraints Two single-coil reluctance actuators are shown in Fig. 1. One is a classical reluctance actuator with a stationary coil and a moving plunger (CLA). A second actuator includes a permanent magnet atop the core (PMB) to allow zero-power latching by means of a passive attraction force [1], [3], [5]. In addition, the actuator height and diameter are 16 and 13 mm, with a stroke of 0.25 mm. Moreover, the plunger of mass 1.2 g experiences an opposing force of 4 to 12N. Finally, the closed-toopen transition can last maximally 4 ms, with a typical valve-open time of several seconds. Open-loop simulation results In an open-loop co-simulation between Simulink and FE software, predefined voltage profiles are applied to the actuators, while the current is limited. In Fig. 2a, the electromagnetic force develops 0.075 ms slower in cases with eddy currents, and the final position is reached 0.115 ms later. This indicates the inherent eddy current damping in the device, slowing down the plunger. In addition, once the movement commences and the airgap closes, the developed electromagnetic force increases rapidly while the opposing force decreases, resulting in a quickly moving plunger. As a result of applying the voltage profiles in Fig. 2b, the corresponding coil currents develop. Note that equal voltages are applied to CLA and PMB until 1.15 ms, after which CLA requires a small hold voltage (1V) to hold the valve open (latch), whereas PMB achieves this passively. Therefore, the hold power can be reduced to zero using PMB. Fig. 2c shows the (in)ability of the actuators to passively latch the valve. The plunger in CLA retracts quickly after (<0.2ms) the supply voltage is removed, as the developed electromagnetic force drops below the opposing force. On the other hand, PMB latches indefinitely, under equal operating conditions, because of the passive attraction force provided by the PM. In general, the predefined voltage profiles produce unnecessarily high forces, indicating that additional control can greatly improve the energy efficiency. Moreover, a significant energy consumption reduction can be achieved by latching passively, and, therefore, reducing to zero the coil current and the hold power using PMB. In addition, plunger closed-to-open movement takes under 0.3ms without achieving a soft landing, while 4ms is allowed. Together, these considerations require to investigate closed-loop feedback control.