Hydraulic powered systems in mobile machines suffer from low operating efficiency during multiple actuator operation owed to large throttling losses when metering the flow to low pressure actuators. State of the art throttling based systems suffer from these losses, demand higher engine power to supply for peaks, couple the prime mover to the dynamic nature of the load and offer no energy recuperation or regeneration. Significant improvement in energy efficiency of mobile machines can be achieved by addressing these drawbacks. This paper presents a novel multi-pressure system aiming to reduce throttling losses, handles power peaks locally, regenerates and stores energy and decouples the supply system from load dynamics. In the multi-pressure system, Proportional bidirectional poppet type 2/2 valves are used to select the best possible pressure line from four levels maintained by integrated hydraulic accumulators while simultaneously adjusting the flow for the actuators, piston and rod sides. A controller is studied and presented for the novel system and the challenges are also discussed. Additionally, possibilities to improve the energy efficiency is also illustrated. Simulation results from a 20-ton wheeled excavator work functions show good controllability and tracking performance as well as regeneration of energy during lowering of the boom. Reduction in energy consumption compared to a traditional load sensing system is also realizable although difficulties defining the novel system's saving abilities is present.
Volumetric flow rate sensors are used in various technical applications. Therefore, it is interesting to use volumetric flow sensors that neither obstruct nor manipulate the flow to be measured nor are restricted to certain flow types and profiles. For this reason, the virtual volumetric flow sensor was developed. A test rig was constructed to validate this soft sensor, which can generate laminar, turbulent, steady, and unsteady flow rates. The dynamic part of the flow is generated by coupling three cylinders and operating a servo valve. In this work, an experimental hydraulic test platform capable of generating reproducible high-frequency oscillatory flow rates is investigated as an enabling system for validating a pressure-based virtual volumetric flow sensor. Two gain-scheduled PID control strategies are implemented to realize the required excitation profiles. Both controllers were intensively investigated on the test rig for various high-frequency scenarios, including pulsations up to 80 Hz. At 80 Hz, the direct controller achieves a normalized mean absolute error of 36.3% (including phase delay). After phase alignment, waveform fidelity corresponds to an nMAE of 10.1%, demonstrating suitability for high-frequency soft-sensor validation. The comparative results show that direct velocity control remains effective up to excitation frequencies of 80 Hz, while indirect position-based control becomes ineffective above approximately 40 Hz due to inherent phase delay. Eventually, the generated dynamic flow rate is utilized to demonstrate the high accuracy of the soft sensor for an oscillation of 20 Hz.
The importance of machine learning (ML) in various engineering disciplines has steadily increased over the past several years, primarily due to the rise in computational power and the development of new, powerful algorithms. ML methods have also found significant applications in fluid power technology, offering substantial benefits such as enhanced system performance, optimized design processes, and improved predictive maintenance. These methods are increasingly used to handle complex, nonlinear systems in fluid power, providing advanced solutions for simulation, system design, control strategies, and condition monitoring. In particular, ML techniques can process vast amounts of data to predict system behavior, identify faults, and optimize energy efficiency, leading to more reliable and efficient fluid power systems. This review aims to introduce and discuss the wide range of ML applications and ongoing research in fluid power, offering engineers a comprehensive overview of the available literature. By doing so, we aim to help engineers identify and select the most suitable and promising ML methods to address their specific tasks and challenges.
Knowledge of the level of entrained air in pump inlets is important for hydraulic system design. Entrained air levels, often quantified as gas void fractions, can change with operation of hydraulic systems, so monitoring entrained air during operations can provide substantial design insight to a hydraulic system. The performance of a SONAR-based entrained air measurement system was investigated on a test hydraulic system utilizing a Coriolis meter as a reference. Two instances of the measurement system were installed on the inlet line of a hydraulic pump in series with a Coriolis meter. One instance of the SONAR-based entrained air measurement system was installed to measure the gas void fraction (GVF) within the flow tubes of a Coriolis flow meter, and a second instance of the SONAR-based entrained air measurement system was installed to measure the GVF within in a section of hydraulic hose in series with the Coriolis meter. Data were recorded to evaluate the SONAR-based entrained air measurement systems both upstream and downstream, as well as collocated with, the Coriolis meter. Speed of sound (SOS) measurements were acquired within the aerated hydraulic oil in the inlet line as the GVF was varied utilizing a variable area flow restriction installed within the pump inlet line. Increasing the restriction reduced the pressure and increased the GVF within the hydraulic fluid due to (1) the existing gas expanding and (2) additional out-gassing from the hydraulic oil. The measured SOS was utilized to determine a gas volume fraction (GVF) within each instance of the SONAR-based GVF measurement system. Additionally, GVF was also calculated from the density measurements of a Coriolis flow meter. GVF measurements from the SOS measurement test section across the Coriolis meter were highly correlated with GVF calculated from Coriolis meter density measurements. The SOS GVFs were shifted up by up to 0.126 % GVF and matched with <0.2% RMSE. For the upstream or downstream SOS GVF measurements, there was variation from those calculated from Coriolis meter density measurements. These differences were anticipated due to pressure differences in the pipe section from that in the Coriolis meter. These pressure differences resulted in differences in the air released from the hydraulic fluid and differing GVFs in the two sections. These results show that the SOS measurement technique can accurately measure the entrained air status of hydraulic systems.
Low-cost fault diagnosis of pneumatic systems has been highly demanded by the industrial community in recent years. In this study, the feasibility of low-cost and low-redundancy fault diagnosis in complex pneumatic systems is investigated by using a minimal number of mature flow and pressure sensors. A pick-and-place demonstration system with 17 pneumatic actuators is taken as the experimental platform. Only one pressure sensor and one flow sensor are utilized to diagnose 132 leakage faults with the help of a one-dimensional convolutional neural network (1D CNN). The average accuracies of leakage fault diagnosis with pressure, flow rate, and exergy data are 89.5%, 96.8%, and 98.9%, respectively. The results are interpreted with Class Activation Mapping (CAM) and Occlusion Sensitivity Analysis (OSA). Overall, it reveals that it is feasible to diagnose multiple faults in complex pneumatic systems with a minimal number of commonly used sensors.
