Engineering topics

Blelloch, Paul A.

Publications and source records attributed to Blelloch, Paul A..

8 records

NASA NTRS2018

Historical Perspective on Fast Coupled Loads Analysis Methods

Spacecraft structural designs are typically verified through a coupled loads analysis (CLA) process, which couples the spacecraft model with the launch vehicle (LV) model to predict low-frequency quasi-static and dynamic responses. The CLA calculations are typically the responsibility of the LV organization, but the spacecraft organization has a vested interest in being able to calculate approximate CLA results during the design of the spacecraft. Because of this, there has long been interest in a method that would allow a spacecraft organization to perform a CLA without access to the full set of LV models and forcing functions. One such method is the Norton-Thevenin Receptance Coupling (NTRC) approach, which is specifically designed to accurately transform LV free accelerations (no payload) into coupled system accelerations (LV plus payload). The purpose of this report is to provide historical context for the NTRC method and compare it with methods that have been used in the past. In particular, it is compared to a frequency-domain substitution method that had been used for a long period of time at the Jet Propulsion Laboratory, and a component-mode-based equivalent to that method.

Blelloch, Paul A.↗

NASA NTRS2018

An Examination of Launch Vehicle Loads Reanalysis Techniques

The typical approach to calculating dynamic launch loads in aerospace applications is coupled loads analysis (CLA). A component mode model of the launch vehicle is coupled with a component mode model of a payload, system modes are calculated, forcing functions are applied, and the dynamic responses are computed. This approach requires the explicit knowledge of the component models for the launch vehicle and payload, as well as the forcing functions. In many situations, the launch vehicle and forcing functions do not change from one analysis to the next only the payload is different. For this type of application, a method called reanalysis was developed to compute the response of a modified payload on the same launch vehicle. If the launch forcing functions are also the same, the approach eliminates the need for the launch vehicle model and the forcing functions and dramatically reduces the computation time. This work investigates the application and accuracy of three previously developed reanalysis methods using a typical launch vehicle and two different payloads. All three methods are based on knowledge of system modes from the original CLA, and Hurty/Craig-Bampton (HCB) models of the original and new payloads. The first method was developed at JPL (Jet Propulsion Lab) and is often referred to as substitution. It is a frequency-domain method, which requires transformation of time signals to and from the frequency domain. The other two methods are time-domain methods that more closely mimic the CLA process. The three methods were applied to two simple examples and a more complex one. The results indicate that the time-domain methods are considerably more robust with respect to modal truncation and other numerical errors than the frequency-domain JPL method.

Reanalysis↗

NASA NTRS1993

A residual flexibility approach to multibody dynamics

Many complex systems can be modeled as a collection of interacting bodies, where the relative motion of the bodies may be large. The dynamics of such systems are simulated using multibody dynamic formulations. Many of these treat each body as a rigid component, but recently the flexibility of the components has been incorporated. This paper presents a residual flexibility formulation of the multibody dynamics problem. The formulation is very simple and offers great computational efficiency since it treats each body as a free structure in space, interacting with other bodies only through interface forces. Each body's accelerations can be solved independently, as can each set of interface forces. We have applied the technique successfully to several special applications, and the initial implementation in a general mechanisms code has given excellent results in comparison to a direct finite element representation of flexibility.

Blelloch, Paul A.↗

NASA NTRS1991

Structural representation for analysis of a controlled structure

The purpose was to determine what reduced order structural representation is most appropriate for coupling with a control system. The goal was to choose a reduced order structural model which retains as closely as possible the characteristics of the closed loop model with a full order structural representation. By characteristics of the closed loop model, it is meant that the closed loop eigenvalues and the closed loop transfer functions from commands to loads and from commands to response. This process does not address the accuracy of the full order model (usually a finite element model) but only the loss of accuracy associated with reducing th model. For the purposes of this study, only collocated sensors and actuators are examined. The choice of a structural representation for noncollocated sensors and actuators is not so clear.

Blelloch, Paul A.↗

NASA NTRS1991

Simulation of the Space Station strut-out condition

A method is presented for reanalyzing a truss structure when one of the truss elements (struts) has failed. The method uses a modal model of the nominal structure coupled with a residual flexibility term to predict the effect of the failed strut without resolving the finite element model. By implementing the method as part of the transient simulation, it is feasible to consider a large number of potential strut failures with a minimum amount of extra effort. Preliminary application of the method to the Space Station indicates excellent agreement with results based on modifying and resolving the finite element model.

Blelloch, Paul A.↗

NASA NTRS1990

Selection of component modes

Structural dynamic models of complex spacecraft are often assembled from a number of component models. In this paper, three methods for reducing the order of these component models are presented by selecting 'important' fixed interface component modes. These methods are applied to two component structures of the Space Station Freedom: a photovoltaic (PV) array and a solar dynamic collector. The reduced order models of these components retain a small fraction of the number of modes in the original models while accurately representing a set of outputs chosen by the user.

Blelloch, Paul A.↗

NASA NTRS1989

Modal representations in control/structure interaction

When control/structure interaction problems are examined, a standard method for representing the structure is to choose a truncated set of normal modes calculated from either a finite-element or a distributed-parameter model. However, the normal modes can neglect important static information about the structure. Using a set of fixed interface modes results in a much more accurate closed-loop model, even when relatively low-bandwidth controllers are used. The fixed interface modes are calculated with control input degrees of freedom held fixed, and standard finite-element software can be used. Illustrative examples include a simple hinged beam and a complex model of the phase-I Space Station configuration.

Blelloch, Paul A.↗

NASA NTRS1989

Modal selection in structural dynamics

An overview of two modal selection procedures for lightly damped structural dynamic models is presented. Both procedures order the modes in terms of their contribution to the input/output dynamics of the model. A complex model of the Phase I Space Station is used to illustrate the application of these procedures to a realistic structure.

Blelloch, Paul A.↗