Authored by: Edited by Leslie Gordon Key points: Resources: |
As CAE software becomes increasingly sophisticated, engineers can now refine designs to their nearly final form. This lets engineers test physical prototypes later in the product-development cycle. The use of CAE software is critical in the space industry, where it’s difficult to create test settings that simulate real-world conditions. Vacuum chambers and wind tunnels help, but they cannot account for all conditions at the same time. That’s why spacecraft producer Thales Alenia Space Italia (TAS-I) in Italy used CAE software in the design, testing, and building of its hypersonic reentry vehicles for the European Space Agency.
Head of aeromechanics and propulsion at TAS-I, Cosimo Chiarelli helped design and test the vehicles. He explains that the physics of atmospheric reentry are complex, so vehicle analysis required a multidisciplinary optimization (MDO) approach to account for all the variables.
Variables include aspects of the spacecraft structure such as geometry (length and shape), as well as the dimensions and material attributes of the shell and thermal-protection system. Other variables include the trajectory (comprising the vehicle’s speed, altitude, and angle of attack), the thermal conditions for the vehicle’s windward, leeward, and nose zones, and the thermal loads the vehicle encounters. The final design accounts for all variables, with a focus on the 150 sec that make up the most-critical portion of reentry.
To improve designs, engineers conducted separate simulations for each of the physics disciplines. They used a collection of software packages and divided the analysis into seven major computational tasks and 40 subtasks, many with their own input and output file types. Engineers used Isight process automation software from Simulia, Providence, R.ŒI., to organize the tasks, manage the execution of TAS-I’s different codes, and aide in the understanding of results from all of the tasks. “The software helped us create flexible simulation workflows and automate the exploration of solutions for the large design matrix,” says Chiarelli.”
To conduct a feasibility study of their new MDO approach, engineers chose a theoretical hypersonic reentry vehicle and applied simplified assumptions. Further streamlining the process, they decided to optimize globally for all variables combined, rather than locally for each individual variable. To minimize costs, engineers applied the process-automation software’s adaptive simulated annealing algorithm, a statistical technique that searches the envelope of design solutions. Isight performed 200 iteration cycles in only a day, assembling several designs that satisfied requirements. Chiarelli says the software handled all the simulations and improved the workflow, helping TAS-I unify its processes and slash analysis times.
After engineers had established a way to efficiently improve designs, they had to verify that the resulting vehicle would survive during the harsh conditions of reentry. The simulation involved a large number of related variables, and engineers again used Isight to manage the workflow.
In past projects, once engineers had defined a trajectory, they identified extreme thermal loads on the design, dividing the procedure into steps. But this approach only let engineers see specific points on the vehicle during worst-case scenarios, so they typically made overly conservative design assumptions, which hurt the craft’s performance.
To apply real-world conditions, engineers plugged a large, proprietary database containing all 25 flightcondition measures into Isight. The goal of the analysis was to assess each time step, for every possible trajectory for every zone on the vehicle. Engineers relied on the software’s capability to combine the separate steps and run the entire process flow.
The software first performed a Monte Carlo simulation that evaluated the model’s performance for the 25 flight parameters during 1,000 sample trajectories. The software postprocessed results for critical variable time histories at 100 different locations on the vehicle for all 1,000 trajectories (a total of 300,000 time histories). Last, it analyzed all the time histories to identify maximum heat loads and fluxes for every point on the vehicle during each trajectory.
As the engineers had hoped, the MDO approach made it easier to calculate loads on the vehicle surface while accounting for inaccuracies affecting the trajectory itself. MDO took only about 48 hr, as compared to the traditional sequential process of design, analysis, and optimization, which would take two weeks, says Chiarelli.
Because of success of the studies, TAS-I engineers now use higherfidelity codes and apply MDO to more complex reentry vehicles. Engineers estimate that man-hours for simulation iterations in these cases could be reduced by about 80% because of fewer manual errors from data transcription, increased design efficiency, and better collaboration between engineering disciplines and departments. Because engineers can address the many design uncertainties of hypersonic reentry, more deep-space probes will return to Earth with valuable payloads.