Robust Position Control of a Four Phase Linear Switched Reluctance Motor for Precision Drug Infusion: Real Flux Linkage Modelling and Load Disturbance Rejection

Precise plunger positioning in infusion pumps is a critical clinical requirement under variable, nonlinear hydraulic loads. This article presents an electrohydraulic model of a four-phase linear switched reluctance motor (LSRM) driving the syringe plunger. A flux-linkage map from finite element analysis captures the electromagnetic behavior, and a velocity-coupled Hagen–Poiseuille model of a clinical fluid path represents the fluidic pathway, so that the flow rate follows from the plunger motion. Control combines a three-loop cascaded structure for position, velocity and current with model-based load feed-forward, phase selection from the co-energy force map and position references derived from each target flow rate. Dynamics are solved with a fourth-order Runge–Kutta integrator, with a realistic encoder model. Nine clinically relevant scenarios span zero to five milliliters per minute, including a mechanical shock applied during acceleration, constant-velocity infusion and holding. Force-sharing commutation is the proposed configuration, and single-phase excitation the reference case. Across the infusion scenarios, the delivered-volume error stays below 0.02 % and the mean flow-rate error below 0.015 %, with a hold oscillation of at most 0.4 µm in the scenarios whose hold window is free of the shock, at the level of the encoder resolution. The residual error is set by the phase commutations: force sharing reduces the commutation error spikes of single-phase excitation from 5.0–8.4 µm to 0.7–1.5 µm, within the ±5 µm clinical band, and the thrust ripple from 68–78 % to about 3 %. The shock, which exceeds the available force margin, displaces the plunger by up to 216 µm; the stroke is recovered within 171 ms and the delivered volume is preserved. Closed-loop control removes the dose errors of open-loop half-stepping, and the trumpet-curve procedure of IEC 60601-2-24, applied in simulation to a time-scaled delivery, gives flow errors within ±0.1 %. This work establishes a reproducible, fully documented simulation reference against which model predictive, adaptive or fuzzy logic controllers can be fairly benchmarked for next generation infusion systems.