This article describes the design method for an electrical machine for a free-piston linear generator. As with other applications for linear machines, defining a steady-state operating point as the basis for the design proved infeasible. To overcome the challenge of optimizing for an alternating motion cycle with accelerations and dynamic load forces, a previously established design methodology was implemented. In addition to the experimental validation of the optimization approach, each research step was presented, beginning with the preliminary choices for the machine topology. This was followed by the optimization of the linear machine and lastly the measurements of its characteristics for comparison with the simulation. A control strategy was determined and a machine prototype constructed that achieved the desired operating cycle, matching the simulated behavior remarkably.
The free piston engine is an Internal Combustion Engine (ICE) without crankshaft. This allows driving an independent trajectory of the piston stroke, which adds a new degree of freedom for the control. This trajectory has to be controlled by using a linear electric machine. Most of the reaseach in free piston engines until now was focused on two-stroke cycles. The present paper focuses on a four-stroke cycle and investigates the particular requirements of the linear machine and its control. Especially in a four-stroke free piston engine it is possible to adjust every stroke trajectory to an optimal one depending on the operating point. Herefore a control structure is proposed and validated on a small experimental prototype of the engine. Unlike rotating engines, the directly driven piston provides also the opportuniy to estimate the instantaneous pressure of the cylinder without additional sensors, which allows to optimize the ignition point among others. For this, an observer is proposed and validated on the prototype.
1 August 2019 in IEEE/ASME Transactions on Mechatronics
This paper describes a novel approach to the design of linear machines that are not run at a steady state operating point. The focus lies on minimizing the power loss to achieve highest possible efficiency. First analytical considerations reveal the potential of designing electrical machines for periodic, accelerated motion. It shows the influence of the machine size on the power losses due to the dependence of machine force as well as moving mass on this parameter. Based on these findings a simulation model was created that can be used with an optimization algorithm. The simulation consists of a stationary and a time dependent finite element analysis of the linear machine in conjunction with an operational model that includes a given movement cycle. It was used to design a flat type linear machine for a free piston linear generator. The simulation results validate the analytical calculations and illustrate the differences to conventional design approaches for linear machines.
A system consisting of an internal combustion engine (ICE) and an electric generator, as used in serial electric hybrid vehicles (EHV) or range-extended electric vehicles, is considered in this paper. The use of a torque controlled generator provides an additional actuating variable for the control of the ICE. In this way, the power of the ICE can be controlled without using throttle actuator while following nearly maximum efficiency trajectory. A simple control algorithm is proposed for this approach based on an inner speed and outer power control loop. The resulting system is simple, robust and efficient requiring less energy storage than the usual on/off control strategy for EHV. The proposed algorithm has been implemented in a prototype and validated experimentally.
A system consisting of an internal combustion engine (ICE) directly coupled to a permanent magnet synchronous machine (PMSM) is considered, which can be used as a range extender in hybrid electric vehicles (HEV) or as a power unit. For control of both, the synchronous machine and the engine, it is necessary to know the position of the rotor / crankshaft. Instead of using a position sensor, the position will be determined indirectly by using only electric variables of the PMSM by a self-sensing method. In general the angle determined in this way is only used for the field orientation and speed control of the PMSM. In this paper it is shown that it can be also used to control the ICE. Therefore, the electric angle has to be transformed to an absolute angle of the ICE relative to the top dead centre (TDC). This angle can be used to set the ignition point, control the timing of the valves and the injection. To transform the electric angle to an absolute angle, the ICE is used as a sensor to determine the TDC. Therefore the compression torque of the ICE is estimated through the PMSM and compared with an ideal model. The resulting benefit of the proposal is that the TDC can be obtained sensorless, with an accuracy of +/- 0.8 degrees.