Open-Loop Control of a Multi-Pressure Rail Hydraulic Excavator

Conventional load-sensing hydraulic excavators provide robust actuator control, but their efficiency is limited by pressure mismatch and throttling losses during multi-actuator operation. Multi-pressure rail hydraulic systems can reduce these losses by providing multiple pressure levels for actuator load matching, but they also introduce challenges in pressure-mode selection, valve-command generation, and mode-transition behavior. This thesis investigates whether a reduced-complexity position-sensorless open-loop local actuator controller can preserve the main energy-saving benefit of a multi-pressure rail hydraulic excavator architecture without relying on a full closed-loop pressure-control or model-predictive-control framework. A 24-ton load-sensing excavator is used as the reference machine, and a dynamic simulation model is developed to evaluate actuator response, rail-pressure behavior, and energy performance.The proposed controller separates force-mode selection from valve-command generation. A Force Index Controller selects the rail-to-chamber force mode, while a valve controller generates the corresponding valve commands without using piston-position feedback as a local motion-control signal. The system is evaluated using property-oriented simulations, a measured 90-degree dig-dump duty cycle, and a controller-level comparison with an MPC benchmark. For the selected duty cycle, the open-loop MPR system reduced normalized pump shaft energy by 16.51% and hydraulic energy by 16.16% relative to the load-sensing baseline. The useful delivered work was slightly different because the realized actuator motion was not perfectly identical to the measured baseline motion; this difference is not interpreted as a productivity improvement. The throttling-loss fraction decreased from 48.85% to 42.16%, indicating that the main energy-saving mechanism is improved actuator-level load matching. Property-oriented simulations showed that mode transitions can introduce local force mismatch and displacement offset, while a diagnostic input-shaping case showed that transition shaping can reduce severe switching offsets. A controller-level comparison with an MPC benchmark under prescribed constant rail pressures is used to examine how actuator-side energy demand, valve-path usage, and pressure-regulation behavior differ between the two controllers. Overall, the results suggest that position-sensorless open-loop MPR control can preserve a meaningful duty-cycle energy benefit, while mode-transition behavior and valve-command generation remain important areas for improvement.