High Performance Computing Terramechanics Methods for Realtime Off-Road Vehicle Mobility Simulation

Overview

Contributors: Justin Madsen, Andrew Seidl, Paul Ayers (University of Tennessee-Knoxville), George Bozdech

In this project terramechanics models were developed to incorporate a physics-based, three dimensional deformable terrain database model with vehicle dynamics mobility simulation software. The vehicle model is contained in Chrono, a research-grade C++ based Application Programming Interface (API) that enables accurate multibody simulations. The terrain database is also contained in a C++ based API, and includes a general tire-terrain interaction model which is modular to allow for any tire model that supports the Standard Tire Interface (STI) to operate on the terrain. Furthermore, the ability to handle arbitrary, three dimensional traction element geometry allows for tracked vehicles (or vehicle hulls) to also interact with the deformable terrain. The governing equations of the terrain are based on a soil compaction model that includes both the propagation of subsoil stresses due to vehicular loads, and the resulting visco-elastic-plastic stress/strain on the affected soil volume. Non-flat, non-homogenous and non-uniform soil densities, rutting, repeated loading and strain hardening effects are all captured in the vehicle mobility response as a result of the general 3-D tire/terrain model developed. Pedo-transfer functions allow for the calculation of the soil mechanics model parameters from existing soil measurements. This terrain model runs at near real-time speed, due to parallel CPU and GPU implementation. Results that exercise the force models developed with the 3-D tire geometry are presented and discussed for a kinematically driven tire and a full vehicle simulation.

Animations

  • Four tires on a dynamic vehicle model with no slip/shear effects on an incline
  • Related Publications

    Madsen, J., Seidl, A., Negrut, D., 2013 "Compaction-Based Deformable Terrain Model as an Interface for Real-Time Vehicle Dynamics Simulations", SAE World Congress April 2013, Detroit, MI, 13M-0241.

    Madsen, J., Seidl, A., Negrut, D., 2013 "Off-Road Vehicle Dynamics Mobility Simulation with a Compaction Based Deformable Terrain Model", ASME IDETC August 2013, Portland, OR, DETC2013-13152.

    A Physics-Based Vehicle/Terrain Interaction Model for Soft Soil Off-Road Vehicle Simulations

    Overview

    Contributors: Justin Madsen, Andrew Seidl, Paul Ayers (University of Tennessee-Knoxville), George Bozdech

    The development of a three-dimensional vehicle/terrain interaction model which is comprised of a tire and deformable terrain model to be used with a real-time vehicle dynamics simulator. The governing equations of both models are physics-based, rather than utilizing popular terramechanics models that are empirical. The tire draws on a lumped-mass model based on a radial spring-damper-mass distribution. The terrain model utilizes Boussinesq and Cerruti soil mechanics equations to determine the pressure distribution and deformation of a volume of soil as a function of vertical and lateral forces applied at the soil surface by the tire. The model treats a column of soil as a system of discretized soil volumes, and the deformation of each is modeled using visco-elasto-plastic compressibility relationships that relate subsoil pressures to a change in bulk density of the soil, which in turn produces soil displacements. Different loading combinations applied by a tire passing over a column of soil will be reflected in the state of each volume of soil contained in the column, rather than treating the column of soil as homogeneous in the vertical direction and only associating one set of parameters with the entire column, e.g. a Bekker type model. Furthermore, the time-dependent elastic and plastic response of the soil to repetitive compression/rebound tire loads is also taken into account. Lateral forces produced by slip and bulldozing effects will also be incorporated into the model. These physics-based models are anticipated to lead to a more realistic vehicle dynamic response when driving on off-road, deformable terrain conditions, especially when repeated loading occurs or non-homogeneous soil conditions are present. Additionally, the changes in soil states can be used to directly compute the energy and power required to deform the terrain. In order to retain the ability to run real-time simulations, a parallel computing approach is considered to leverage the inherently parallel nature of performing multiple independent terrain geometry queries and soil-mechanics calculations. Numerical experiments are run on a single soil volume node under a known load and for a simplified tire model applying normal forces on the surface of the terrain. Results are given for the vertical plastic soil deformation, and for the power and energy required to perform the deformation.

    Animations

  • Lugged rigid tire on Deformable terrain, rear view
  • Lugged rigid tire on Deformable terrain, top view
  • Related Publications

    Madsen, J., Negrut, D., Reid, A., Seidl, A., Ayers, P., Bozdech, G., Freeman, J., and O'Kins, J., 2012, "A Physics-Based Vehicle/Terrain Interaction Model for Soft Soil Off-Road Vehicle Simulations," SAE International Journal of Commercial Vehicles, 5(1), pp. 280-290.

    Effects of Higher Order Blade Modes on Aeroelastic Wind Turbine Structural Analysis

    Overview

    Contributors: Justin Madsen, Brian Resor, Daniel Laird (Sandia National Labs)

    The purpose of a modern wind turbine is to generate the greatest amount of electricity possible while keeping production and maintenance costs low, leading to the lowest possible cost of energy. Power is generated by the wind interacting with the turbine blades, which produces large aerodynamic lift forces on the blade and rotor structures. Current trends have favored increasingly larger turbine blades in order to reduce the cost of energy and allow efficient operation in offshore locations. Large blades present a design challenge not only due to the higher loads transmitted, but also because they tend to be more flexible and gravity effects are more significant. The effect on the structural dynamics is such that as bend-twist coupling in the blade design or upsizing of turbine size can excite phenomena such classical aeroelastic flutter at or near normal operating conditions. A key aspect of this work is to identify the operating conditions which excite aeroelastic instabilities in a multi-megawatt size wind turbine model. These results will be combined with previous work to predict at what turbine size or blade design type will excite aeroelastic instabilities at or near normal operating conditions. This work will also addresses the effects higher order blade modes have on performance from the turbine design analysis standpoint and aims to expose the operating conditions that necessitate including these modes when evaluating a specific turbine design. A 5MW wind turbine model is utilized for a majority of the experimental simulations.

    Animations

  • Aeroelastic Turbine in ADAMS with AeroDyn
  • Detailed Overview

    Recent trends in the wind energy industry have led to wind turbines that are increasingly large in size with peak outputs typically greater than 1 MW. Theses turbines rely on blades whose length varies from 35m for a 1.5 MW turbine to greater than 60m for a 5 MW machine. As this trend continues, tools for analysis of the aeroelastic system response will need to be applied accordingly to account for the additional physics and phenomena associated with larger turbines and blades. As the blade growth trend continues it changes the structural dynamic response of the system in that the mass of a typical wind turbine blade scales with the cube of the radius, i.e. R3. At the same time blades are manufactured to be more flexible which leads to lower natural frequencies for structural modes and thus higher order modes are excited more easily under multiple operating conditions. It is this trend of higher order structural modes occurring at lower natural frequencies that motivates this work. Engineers tasked with the analysis of a wind turbine design would likely prefer using an aeroelastic simulation tool that can successfully capture the physical phenomena that impact the operation of the turbine while being as computationally efficient as possible in order to reduce the time required for analysis. For example, The National Renewable Energy Laboratory (NREL) created and actively maintains the FAST software and its source code, which is freely available to the public. FAST is a medium-complexity code for nonlinear aero-servo-elastic analysis of horizontal-axis wind turbines. It can also extract linear state-space models for controls design and can be used to generate MSC.ADAMS models. The FAST code models the wind turbine as a combination of rigid and flexible bodies. Aerodynamic forces are applied to the blade of the FAST model using AeroDyn, another NREL software offering. FAST with AeroDyn was evaluated by Germanischer Lloyd (GL) WindEnergie and found suitable for the calculation of onshore wind turbine loads for design and certification. The blades are modeled using modal superposition of the first two flap-wise modes with the first edgewise mode. It is a computationally efficient code but lacks the ability to capture non-linear blade deflections and torsional blade modes that are likely to be present in larger blades.

    As blades become long and slender, effects such as bend-twist coupling can lead to detrimental instabilities such as flutter. Also, when blades experience twist it modifies the angle of attack, impacting the lift and drag coefficients along the length of the blade. Both these phenomena can have an effect on performance and life of the turbine. On the other hand, blades can be intentionally designed to twist as they approach their rated operating speed to change the angle of attack and mitigate loads. In either case, the inclusion of non-linear blade deformations and higher order structural modes becomes important and require high-fidelity aeroelastic models to analyze. MSC Corporation markets a multibody analysis software analysis package called MSC.ADAMS, or ADAMS. FAST can be used as a preprocessor to create ADAMS wind turbine models. ADAMS representations of wind turbine models are of much higher structural fidelity than their FAST counterparts due to the blades being represented as a series of lumped masses connected with Timoshenko type beam elements. This allows for the inclusion of large blade deformations as well as higher order modes such as torsion. Beam parameters are extracted from FE models of blades using the Beam Property Extraction tool developed at Sandia National Labs. Similarly to FAST, the aerodynamic forces are applied to the ADAMS wind turbine model using the AeroDyn subroutine. ADAMS, together with FAST, has been given certification by GL for wind turbine design calculations. By using both FAST and ADAMS aeroelastic simulation codes, nearly identical turbine models can be created and simulated using the same wind input files from AeroDyn. The only difference is the method of representing the flexible tower and blades, where FAST uses the computationally efficient modal superposition approach and ADAMS utilizes a higher fidelity lumped mass approach.

    Comparing results from various simulations that model actual turbine operating conditions, higher fidelity aeroelastic models are required when higher order blade structural modes impact the turbine performance. Here, both FAST and ADAMS simulations are run where the only difference between the models will be the representation of the flexible tower and blade structures. The wind and turbine conditions listed in IEC 61400-1 Ed. 3 will be the basis for running the comparison simulations. The goal is to identify what, if any, load cases necessitate the use of a model that includes higher order structural modes to properly capture the dynamics involved when simulating a 5MW turbine.

    REFERENCES

    1. J Jonkman, S.B., W. Musial and G. Scott, Definition of a 5-MW Reference Wind Turbine for Offshore System Development, 2009, NREL: Golden, CO. 2. P.S. Veers, T.A., et. al., Trends in the Design, Manufacture and Evaluation of Wind Turbine Blades. 2003. 6(3). 3. F. Rasmussen, e.a., Present Status of Aeroelasticity of Wind Turbines. 2003. 6(3). 4. Jonkman, J., NWTC Design Codes (FAST), 2010, http://wind/nrel.gov/designcodes/simulators/fast: Golden, CO. 5. B.J. Jonkman, J.M.J., Documentation of updates to FAST, A2AD, and AeroDyn, 2010, NREL. 6. Laino, D., NWTC Design Codes (AeroDyn), 2010, NREL: Golden, CO. 7. Hansen, M.H., Aeroelastic Instability Problems for Wind Turbines. 2007. 10(6). 8. D.J. Malcolm, D.L.L., Extraction of Equivalent Beam Properties from Blade Models. 2007. 10(2). 9. J. Jonkman, D.L., NWTC Design Codes (ADAMS2AD), 2010, NREL: Golden, CO. 10. Lobitz, D., Aeroelastic Stability Predictions for a MW-sized Blade. 2004. 7.

    A Stochastic Approach to Integrated Vehicle Reliability Prediction

    Overview

    Contributors: Justin Madsen, Dr. Dan Ghiocel

    This research addresses some aspects of an on-going multiyear project of GP Technologies in collaboration with University of Wisconsin-Madison for US Army TARDEC. The focus of this research project is to enhance the overall vehicle reliability prediction process. A combination of stochastic models for both the vehicle and operational environment are utilized to determine the range of the system dynamic response. These dynamic results are used as inputs into a finite element analysis of stresses on subsystem components. Finally, resulting stresses are used for damage modeling and life and reliability predictions. The integrated approach combines the computational stochastic mechanics predictions with available statistical experimental databases for assessing vehicle system reliability. Such an integrated reliability prediction approach represents an essential part of an intelligent virtual prototyping environment for ground vehicle design and testing.

    Detailed Overview

  • See Justin Madsen Master's Thesis
  • Animations

  • High-speed
  • Wide turns
  • Rolling hills
  • Related Publications

    Madsen, J., Ghiocel, D., Lamb, D. Gorsich, D., Negrut, D. 2009, "A Stochastic Approach to Integrated Vehicle Reliability Prediction", ASME IDETC/MSNDC.


    Electric Mining Shovel Simulation using a Co-Simulation environment to compare efficiency of AC/DC motors

    Overview

    Contributors: Justin Madsen, Martin Tupy

    The goal of this project was to determine how the dig cycle time of a large above ground electric mining shovel was affected by varying the types of electric motors that actuate the shovel. The implementation was carried out by lab members Justin Madsen and Martin Tupy. An accurate dynamics model of the shovel was created in MSC/ADAMS, and the corresponding motors and controls were modeled using Simulink. Through a co-simulation environment, the lab members were able to successfully create an accurate simulation of the shovel on the system level. Dig cycle times were representative of actual measured values, and the simulations were able to show

    Animations

  • Adams/Simulink co-simulation of dig cycle