Four-Quadrant Force Sharing Control of Switched Reluctance Generator for the Application of Optimal Wave Energy Conversion

Linear switched reluctance generators (LSRG) are promising candidates for direct-drive wave energy converters (WECs) due to their low cost and high reliability. The generator should be properly controlled to actively change the wave-body interactions to improve the wave energy capture efficiency. It requires a two-layer, cascaded control scheme: an upper layer is the energy-optimizing controller providing a force command, and a lower layer is the generator controller tracking this command. Because the WEC oscillates with waves, its velocity is bi-directional. Meanwhile, the optimal force is usually not strictly passive but contains reactive components. Consequently, the LSRG should be controlled to track the desired force in all four quadrants. However, existing LSRGs for WECs typically operate with a fixed on-off angle, limiting their ability for energy-optimizing control. Much research focuses on torque (force) tracking with ripple minimization with constant force and constant speed for switched reluctance motors. There is a gap in experimental testing under WEC-specific operation conditions. This paper proposes a four-quadrant force-sharing strategy for LSRGs for WECs. First, two classic energy-optimizing control strategies, i.e., damping control and reactive control, are introduced, with output force commands linear to the body’s velocity and displacement. Next, the basic model of LSRG is derived, and an exponential-sinusoidal model describing the position-current-force characteristics is adopted. Two basic control methods of LSRG, the constant-current control and the sinusoidal force-sharing control, are introduced. The constant-current control maintains a fixed phase current, resulting in a fluctuating force due to the nonlinear force characteristics. The sinusoidal force-sharing control distributes the total force command into two phases according to a sinusoidal function. Ideally, this produces a constant total force. However, the DC voltage is limited, and an arbitrary current command cannot be tracked in time, so the total force also contains ripples. Therefore, an optimized force-sharing function is proposed to minimize current loss within the constraints of DC voltage and total force command. A discrete force-sharing function can be solved, stored in the controller, and implemented in real-time. Finally, a force distribution table for all four quadrants is derived for implementation. An experimental platform simulating WEC operation is established, with an LSRG as the generator and a permanent-magnet synchronous motor (PMSM) as the prime mover. The parameters of the LSRG, especially the position-current-force characteristics, are identified through fixed-position excitation experiments. Then, three tests are conducted: (1) the classic, constant-velocity test, (2) the WEC’s damping control test, and (3) the WEC’s reactive control test. For each test, constant current control, sinusoidal force sharing, and optimized force sharing are compared in terms of total force. It is verified that under all the conditions, the optimized force-sharing strategy outperforms the constant-current control and sinusoidal force-sharing control, achieving the desired force with minimal ripples. The proposed method meets the energy-optimizing control requirements for WECs, as the changing rate of current profiles is constrained by considering DC voltage. © 2024 China Machine Press. All rights reserved.