Energy saving is always a topical issue in the off-highway industry due to increasing fuel costs, and particularly with earthmoving machines (EMM), this aspect is amplified by more stringent emissions regulations that impact off-highway vehicle development. These considerations dictate a continuous strive toward improvements and more efficient solutions. To accomplish such objectives, a strong reduction of hydraulic losses and better control strategies of hydraulic systems are needed.
An efficient way to analyze these issues and identify the best hydraulic solution is through virtual simulations instead of prototyping. To build a hydraulic excavator virtual model, however, some problems relative to different aspects arise. For instance, loads on actuators (both linear and rotational) are not constant, and pumps are driven by a real engine whose speed depends on required torque.
Furthermore, the level of detail used to simulate each component of the circuit is extremely important. The greater the detail, the better the machine’s behavior can be portrayed, but this obviously entails a heavy impact on simulation time. Therefore, engineers must decide the model level of detail to acquire all the phenomena deemed necessary for a correct evaluation of machine performance.
Many researchers have worked on the problem of presenting hydraulic models of EMMs but always exercised restrictive hypotheses in the course of their studies. For example, some do not treat pump modeling, preferring to address proportional directional control valves (PDCVs) and arm kinematics; others concentrate on pump and kinematics modeling, leaving aside the PDCVs design. The influence of hoses and actuators cushioning is often neglected as well as that associated with various hydraulic resistances. Moreover, attention is limited to the arm circuit, forgetting that excavators have other users such as turret swing and travel that are as relevant as the arm circuit.
Following such considerations, researchers from Politecnico di Torino have determined the necessary steps to develop a detailed hydraulic circuit model of a commercial compact excavator. In particular, the entire hydraulic circuit was modeled within the AMESim environment by using both standard and detailed libraries to better describe the circuit. Their work also made use of simplification hypotheses. For instance, modeling of pilot stage valves was not considered, interaction between track and soil was omitted, and turret inertia was held constant during simultaneous movements. However, complex aspects were carefully taken into account, considering the specific application such as modeling the arm’s geometry.
Various approaches were investigated. Some used the Planar Mechanical Library (PLM) application within AMESim, while others made use of coupled simulation between Virtual Lab Motion (VLM) and AMESim, where VLM handled the body parts of the excavator. In general, coupled simulation performs well and grants easy settings of parameters but is expensive in terms of simulation time.
Wishing to limit simulation time, the PLM was chosen to model the arm. In addition, PDCVs behavior, mainly in case of flow saturation, must be considered and properly managed to avoid the loss of controllability. This led to an in-depth modeling of PDCVs and their pressure compensators. As to the pump model, a compromise solution between quality of results and simulation time was achieved.
In fact, attention focused on displacement control devices, simplifying parts relative to flow rate generation. Other valves such as those present in the travel and turret swing were modeled with the Hydraulic Component Design Library. The internal-combustion engine accounted for speed variation with torque, and the evaluation of fuel consumption to rate the quality of the hydraulic solutions was considered using a fuel consumption map.
Overall, the researchers focused on the analysis and modeling of the whole hydraulic circuit that, beside a load sensing variable displacement pump, featured a stack of nine PDCV modules of which seven were of the load sensing type. Loads being sensed were the boom swing, boom, stick and bucket, right and left track motors, and work tools. The blade and the turret swing did not contribute to the load-sensing signal.
Of specific interest were the peculiarities that were observed in the stack. In fact, to develop an accurate AMESim model, the stack was dismantled and all modules analyzed and represented in a CAD environment as 3-D parts. The load-sensing flow generation unit was replaced on the vehicle by another one whose analysis and modeling were developed using available design and experimental data.
Although both the Hydraulic Component Design Library as well as the Planar Mechanical Library were used extensively in the process of modeling the entire circuit, some simplifications became necessary. As the modeling phase was developing, a number of field experiments on the vehicle were also performed, which served the purpose of providing reference data to the end of progressing with the validation phase of the AMESim model.
Regarding predictive simulation results, those were appropriately consistent with gathered experimental outcomes. The AMESim model will be instrumental for upcoming analyses that foresee the substitution of the original stack with other load-sensing post-compensated modules so to assess with reasonable confidence the effects on potential energy savings.
Future steps are aimed at further modifying the hydraulic circuit by replacing the PDCVs stack to predict potential energy savings.
This article is based on SAE International technical paper 2012-01-204 by Gabriele Altare, Damiano Padovani, and Nicola Nervegna, Politecnico di Torino.