Products
Nightsky Systems, Inc. was founded to develop and market the Disturbance-Optics-Controls-Structures (DOCS®) Toolbox to the aerospace and other communities, and to extend the tool with new capabilities to support existing and future applications.
DOCS
The Disturbance-Optics-Controls-Structures (DOCS) Toolbox is a MATLAB® -based modeling and design package for performing Integrated Modeling analysis and optimization. DOCS was developed under a Phase II SBIR for NASA-Goddard. It consists of an integrated system of MATLAB routines that provide for the definition, solution, and documentation of a dynamics optimization problem for complex opto-mechanical controlled structures.
What is DOCS?
DOCS is a framework for defining dynamic performance, for example RMS wavefront error for an imaging system, as an explicit function of structural, control system, disturbance, and optical parameters. The performance functional can then be analyzed using a variety of techniques in order to map out the design space and optimize the system performance.
Parameter dependent modeling
DOCS uses an innovative technique for computing the variation in dynamics as a function of component design parameters such as structural stiffnesses and masses, control system gains and dynamics, disturbance mechanisms, and optical sensitivities. The system dynamics can be rapidly and efficiently re-computed for any combination of parameter values, making it possible to completely map out the performance space as a function of the design parameters. In particular, the technique allows for the continuous variation of structural design parameters such as moduli, masses, areas, and so on, using only a single Normal Modes solution sequence.
Furthermore, the analytical gradients of the closed loop dynamics with respect to any parameters can be computed in closed form, enabling a wide range of gradient-based optimization and analysis techniques that would otherwise be infeasible using finite difference gradient approximations.
Model conditioning
DOCS supplies a suite of tools for integrating and numerically conditioning analysis models to maintain prediction accuracy. Model units can be specified, queried, and changed. Several model reduction tools are supplied, for application to the structural model prior to integration (preserving the Normal Modes state basis) and to very high order arbitrary basis integrated models, using a stabilized balanced reduction algorithm that can accommodate systems of 4000+ states.
Dynamic performance
DOCS comes with a wide range of objective functions for typical dynamic analyses: peak response, RMS and RSS response, gain and phase margins, and many others, computed with a state space realization, in the time domain, or in the frequency domain. Analytical gradients for all functions are supported. The routines use low level computational routines optimized for structural systems. The routines can be 30-50 times faster than standard tools for typical optomechanical systems, enabling a variety of analyses that would otherwise be computationally intractable.
Analysis and redesign
The true power of the DOCS framework is realized when the objective function and gradients are used to support system redesign. A series of innovative analysis routines are used to sequentially map the design space and to identify a robust design that is cost and risk optimal:
- critical modes analysis: the contributions of all structural mode to the response are ranked, automatically generating a list of modes that dominate the response.
- sensitivity analysis: the sensitivity (gradient) of the performance to any design parameter can be analytically computed. Parameters with the largest sensitivities have the highest influence on the cost, and are good candidates for targeted redesign efforts. Modal parameter sensitivities (to frequency, damping, and residue) determine which modes need to be held to a tight frequency tolerance. Structural parameter sensitivities (material modulus and density, element area and bending inertia, lumped stiffness and mass) indicate which structural design parameters have the highest influence on the dynamic performance.
- design space exploration: the performance can be mapped out on a parameter grid one parameter at a time ( univariate analysis) or two parameters at a time ( bivariate analysis). The resulting cost versus parameter maps give a visual representaion of the shape of the response curve (monotonic or multimodal, and range of variation over the allowable parameter space). For one- or two-dimensional parameter spaces, variate analyses can be used to determine the optimal design. For higher-dimensional problem spaces, the coupling between parameters must be quantified to identify strong interactions that cannot be independently optimized. A factorial analysis can be used to investigate an arbitrary number of parameters across a coarse grid. Algorithms such as Yates' method are used to generate a table of the main, two-, and three-factor effects.
- isoperformance: for a problem with more parameters than metrics, there is generally a non-unique set of parameters that will meet a given performance requirement. Isoperformance analysis is a technique for generating a list of all parameter sets that result in a target value for the (scalar or vector) objective. The results can be used in systems trades to determine the optimal parameter set based on non-performance objectives such as cost or cost risk.
- uncertainty analysis: uncertainties or tolerances in design input quantities such as moduli lead to uncertainties or error bars on the performance metrics. DOCS includes a range of techniques for evaluating performance uncertainty, from approximate but fast Linear Covariance Propagation that models the parameter as Gaussian and the performance function as linear, to slower techniques that map an arbitrary input Probability Density Function (PDF) to the exact PDF of the performance metric. The analysis results can be used to redesign the system for improved robustness, or to place requirements on the tolerances for design input quantities.
- optimization: using the DOCS analytical sensitivities, simultaneous optimization of structural and control system parameters on the fully integrated disturbance-to-performance process model can be efficiently performed.
Together, the DOCS analysis tools form a design optimization system that can efficiently solve present and future controlled-structure dynamics problems that are beyond the capabilities of standard design approaches.
Outputs and documentation
DOCS is designed to generate tabular, graphical, and binary output data in common formats. Graphical and tabular data can be exported directly to Microsoft® PowerPoint® presentations. Optimized NASTRAN structural designs can be exported as bulk data decks.
Who uses DOCS?
DOCS is currently employed by NASA Goddard Space Flight Center and the Jet Propulsion Laboratory on numerous missions including the James Webb Space Telescope, Terrestrial Planet Finder (both Coronagraph and Interferometer), Space Interferometry Mission, Solar Dynamics Observatory, and Fourier-Kelvin Stellar Interferometer. It is also employed by the Association of Universities for Research in Astronomy for the Giant Segmented Mirror Telescope and Thirty Meter Telescope.
DOCS grew out of, and has been applied primarily to the space-based optical observing community. However, it is a general purpose modeling tool that can support a wide range of applications. Please contact us to find out how we can help with your design activities.
Pricing
Please contact us for pricing information.
