About 25% of space-borne electro-optical (EO) sensor programs in both the civil and National Security Space (NSS) communities have experienced reduced on-orbit reliability, as well as cost and schedule overruns of 100% or more1-3. Many of these EO sensor program over-runs can be largely attributed to delays, errors, and inadequate communication that occur at the many handoff points between team members and contractors in the current design process. This leads to the late discovery of technical problems, making them more expensive and time-consuming to fix.
The design of EO sensors requires careful attention to the thermal and structural effects that adversely affect optical performance in terms of instrument pointing accuracy and image quality. Teams of domain experts, each focusing on a separate aspect of the sensor, work towards understanding and managing its complex behavior. The current engineering process is typically fragmented into silos of experts, tools and data. Separate models are constructed, one for each of the disciplines (mechanical CAD, thermal, structures, and optics). Analysis proceeds within each discipline silo with relatively infrequent interaction between groups. There is no organized way for individual domains, or the team as a whole, to manage their design and analysis models over the entire project.
Project design reviews are typically conducted using time-consuming “static” reports and PowerPoint summaries. There is no unified view of the engineering models that represent the product’s behavior, nor is there easy access to key system performance data.
An engineering project team at The Aerospace Corporation comprised of a lead structural, thermal, optical, and mechanical design engineer, directed by Senior Project Engineer Dr. David Thomas, has successfully implemented a new collaborative systems engineering approach on an actual flight hardware program, reducing each design evaluation cycle by over 50%, while providing better insight into the multi-disciplinary behavior of a space-borne sensor4. Using expertise captured in bi-directionally associative design/simulation process templates, integrated structural, thermal, optical analyses were performed in a matter of days vs. weeks/months in the traditional “silo” approach. The Aerospace Corp. has also recently participated in the contractor’s final thermal vacuum (TVAC) testing of the flight hardware to correlate and validate the STOP (structural-thermal-optical performance) process models and results. A significant reduction in the cycle time of the EO sensor design evaluation and validation was achieved while meeting overall sensor design reliability and optical system performance target levels.
The Aerospace STOP Project
The Aerospace team used this new approach on an independent STOP analysis of a critical lens subassembly in support of an instrument contractor’s final TVAC testing. A high fidelity STOP model was developed to calculate visible channel focus shifts and image quality impacts due to thermally-induced structural deformations and refractive index changes (Figure 1) to conduct an assessment of the contractor’s focus control method.
The design concept for the optical lens subassembly employed active heaters which were applied to the outer surface of the lens retainers (Figure 2). Three different on-orbit thermal equilibrium conditions were simulated in the STOP analysis: Hot, Nominal, and Cold Soak conditions.
The analysis details were captured in a system-level design/simulation template as well as in the individual discipline templates, allowing the team to easily run a large number of STOP analysis design studies while modifying both geometric and non-geometric model parameters, and the thermal environment conditions. After each STOP analysis iteration, key performance metrics were available for immediate visualization in the Project Dashboard and compared against the system analysis performance requirements.
The team performed STOP analyses using Comet’s Performance Engineering Workspace, combined with their current commercial CAD and CAE tools — Pro/ENGINEER®, Thermal Desktop®, Nastran, Abaqus®, SigFit, CODE V®, Excel® and Matlab®. The workspace provided a single, consistent view of all the data — models, environments, processes and results — allowing the team to create, share and access data easily. The workspace allowed domain experts to work within their own “domain sandboxes” to understand their aspect of the sensor, but also to work together with other domain experts to gain overall system performance insights. By automatically reusing the expertise captured by the experts in simulation templates, manual data handoff errors were eliminated and the confidence in the accuracy of each analysis iteration was significantly increased.
Throughout this project, the team worked independently in their offices, as well as in concurrent sessions held in a common location, always sharing and managing design and simulation data within a single project tree data model located on a secure, shared disk. The keys to success were easy and constant access to system performance metrics, regardless of the tool that generated the data, and team collaboration throughout the entire multi-physics design process. By working within the integrated work-space, each team member quickly and easily saw how their changes affected other aspects by running the template and reviewing the results in the project dashboard. This increased the systems knowledge of the entire team, contributing to a better general understanding of the sensor’s physical behavior.
STOP Process Template Execution
Comet’s Performance Engineering Workspace allowed the domain experts to capture and automate various analysis processes across multiple physics domains. The following outline is a simplified overview of the various tasks automated in the STOP process:
- The STOP process begins with importing a “tagged” 3-D Pro/ENGINEER® model of the instrument geometry into the workspace.
- Independent thermal and structural meshes are automatically generated, based upon rules specified by the respective analysis domain experts.
- The coarse thermal mesh is then formatted and input into Thermal Desktop® (C&R Technologies) to perform the thermal analysis and the temperature fields are mapped to the higher density structural mesh.
- Thermally-induced structural deformations are then evaluated in Abaqus® (Dassault Systemes SIMULIA®), accounting for complex contact stresses between the “floating” lenses and rings.
- The resulting thermal and structural results act as inputs to Sigmadyne’s SigFit code which computes the perturbations to the optical system due to refractive index changes (dn/dt) in the lens components and the deformations of the optical surfaces.
- The modified optical prescription is then formatted for execution in CODE V® (Optical Research Associates) for evaluation of optical performance. CODE V® calculated the optical behavior of the critical lens assembly using the perturbed model.
- Key performance metrics were immediately displayed in the Comet Dashboard and directly compared against system requirements as each STOP analysis was completed.
Typical CAD or CAE software templates that capture analysis processes are specified using a particular version of the geometry. These can be reused with only small changes to the geometry. The key technology of the Comet workspace is the Abstract Engineering Model (AEMTM), a single integrated data model that captures all the CAD and CAE data at all levels of model fidelity, in a manner that is independent of the underlying tools. The AEMTM allows the STOP template to be reused across widely varying geometry changes, with little or no data re-entry, significantly reducing the rework that is usually required. As the geometry changes, the new CAD models are re-imported and the abstract model automatically reattaches all the engineering data to the new version of the model.
As the AEMTM is component-centric and not geometry-centric, analysts can also create analysis processes that perform calculations on models at any desired mixed level of fidelity. The AEMTM spans the traditional chasm between low fidelity modeling (without CAD geometry) and high fidelity modeling; results from low fidelity calculations can be fed to downstream high fidelity calculations and vice versa. The AEMTM also allows users to create and manage multiple representations of each component in the product structure; these representations are required for various types of downstream analyses, different physics calculations at different levels of fidelity. For example, the optics representation of a lens element and the 3-D CAD representation of the same component are managed simultaneously; the former is used to perform optics calculations while the latter is used for creating both the thermal and structural meshes.
From the STOP analysis results, it was discovered that, despite high axial thermal gradients and smaller radial thermal gradients across the lenses, the focus shifts due to deformations at each end of the assembly cancelled each other out for the thermal soak cases that were analyzed. This was a highly non-intuitive and unexpected finding that provided insight into the behavior of the sensor. The finding was that the unconventional, active thermal controls in the EO sensor design, using two heaters that had the same power settings applied over significantly different lens mounting surface areas, were actually adequate to ensure good optical performance in the field4.
For the TVAC test correlation activity, the STOP process template automatically extracted the thermal results from Thermal Desktop® at test configuration thermocouple locations and plotted these against the actual TVAC test data. The team directly used such test data, stored within an Excel spreadsheet, to validate the high fidelity simulation models. It is important to note that the workspace does not generate the underlying calculations within each physics domain or at the overall system level. That is all still performed by individual solver codes so the accuracy of any simulation iteration is still very much dependent on the underlying modeling and analysis assumptions made by the individual domain experts, just as it is done manually today.
The STOP Project Results
Results of the STOP project were as follows:
- The Aerospace team implemented a new collaborative systems engineering approach on an actual flight hardware program, reducing overall design evaluation cycle time by over 50%. After the STOP template was fully developed and test validated, each new STOP analysis iteration could be fully performed and evaluated within a single day.
- The team conducted real-time design reviews with program management and the customer directly within the integrated workspace. All the key performance data and full 3-D models and results were available for these discussions which centered around system-level characteristics impacting the sensor behavior and performance.
- The integrated analysis approach provided superior physical insight into how the thermal control approach works and, hence, how to further improve the small residual focus errors that remain.
- The team eliminated most of the manual rework that usually accompanies changes to the 3-D CAD geometry, which saved significant time and eliminated human errors.
- The integrated environment and concurrent engineering approach enabled each team member to develop a deeper understanding of the multi-disciplinary behavior of the overall system.
- The project was a successful assessment of concurrent engineering practices where engineers used existing analysis tools, making the transition to a new approach easier.
- Engineers received instant feedback on how accurately the STOP model was predicting the system design performance and quickly adjusted individual domain models to increase the accuracy of subsequent design iterations.
According to Dr. Thomas, “Further work is planned to support On-Orbit Testing (OOT) of this payload and final thermal vacuum testing of a second pay-load. Our baseline optical design model will be updated to include fabrication, alignment, and gravity induced effects, and more of the visible channel components will be added to the model to allow higher fidelity comparisons of predicted and measured visible channel image quality. We also plan to add an adapter to the Comet environment so that the controls algorithm software may be included in the integrated analysis. This will allow the entire focus control system to be modeled from ground command through final image quality.”
- Pavlica, Steve and William Tosney, “Assessment of NRO Satellite Development Practices,” The Aerospace Corporation, 2003.
- Young, Thomas A., Chair, “Report of the Defense Science Board/Air Force Scientific Advisory Board Joint Task Force on Acquisition of National Security Space Programs,” Office of the Under Secretary of Defense for Acquisition, Technology, and Logistics, May 2003.
- John Leon and Juan Rivera, chairs, “NASA Instrument Capability Study Final Report,” NASA Office of the Chief Engineer, NASA Headquarters, Washington, DC, December, 2008.
- Jason Geis, Jeff Lang, Leslie Peterson, Francisco Roybal, David Thomas, “Collaborative design and analysis of Electro-Optical sensors,” Proceedings of the SPIE Optics+Photonics, 4 August 2009.