Model predictive estimation of evolving faults

Samar, S.; Gorinevsky, D.

doi:10.1109/acc.2006.1656568

Cited by 5 publications

(4 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…where I andM G are given by (8),F T HRU ST ,F AERO ,M T HRU ST ,M AERO are given by (3), (4), (9), and (10). By substituting accelerationsẇ from (13), (2), (7) and the map g(w, q; X = 0) from (18) into (17) we obtain the residuals to ber…”

Section: Prediction Residuals For Estimationmentioning

confidence: 99%

“…This is discussed and demonstrated in several aerospace-related applications in. 9,11,[17][18][19] Unfortunately, for the application in hand this approach cannot be applied as a direct extension of the linear models considered so far in this paper. The reason is that we do not know the sign of the components of forces and moments and whether they are decreasing or increasing in advance before the case breach location is known.…”

Section: Ivc Extensionsmentioning

confidence: 99%

See 1 more Smart Citation

Early Detection of Solid Rocket Motor Case-Breach Failures

Gorinevsky

Mah

Timuçin

2007

AIAA Guidance, Navigation and Control Conference and Exhibit

View full text Add to dashboard Cite

This paper describes work on early detection of space launch vehicle failures that could lead to loss of control during ascent. The particular focus is on the Solid Rocket Motor (SRM) comprising First Stage of the Crew Launch Vehicle (CLV). Some of the most important failure modes leading to loss of vehicle control are caused by hot gases ejected from the SRM combustion chamber through a case breach, through a joint between SRM segments, through a nozzle joint, or through an igniter seal. These hot gases create lateral thrust augmentation, with the resulting tilting moment possibly leading to loss of control. By using Thrust Vector Control (TVC) gimballing of the main nozzle, the flight control system tries to counter any attitude disturbance, which could mask the moment augmentation till it overwhelms the TVC and the control authority is lost. The approach proposed and demonstrated in a medium-fidelity simulation of the rocket dynamics employs continuous model-based multivariable estimation and monitoring of the augmentation forces and moments for early detection of the failure. The computations are performed in two stages. First, a 6-degree-of-freedom model of vehicle dynamics is used to compute the predicted vehicle accelerations. Second, the differences with the observed accelerationsmodel prediction residuals-are used in optimal multivariable algorithms to estimate the thrust augmentation. A case breach failure may thus be detected well before its effect on trajectory becomes noticeable in the raw data.

show abstract

Section: Prediction Residuals For Estimationmentioning

confidence: 99%

Section: Ivc Extensionsmentioning

confidence: 99%

Early Detection of Solid Rocket Motor Case-Breach Failures

Gorinevsky

Mah

Timuçin

2007

AIAA Guidance, Navigation and Control Conference and Exhibit

View full text Add to dashboard Cite

show abstract

“…Estimation with the monotonic walk model was considered in our earlier papers [5], [16], [15]. The MAP batch estimation problem was formulated as a Quadratic Programming (QP) with linear monotonicity constraints.…”

Section: Introductionmentioning

confidence: 99%

Efficient filtering using monotonic walk model

Gorinevsky

2008

2008 American Control Conference

View full text Add to dashboard Cite

Abstract-This paper proposes a nonlinear filter for estimating monotonic underlying trend from noisy observations. The filter computes Maximum A posteriori Probability (MAP) estimate using a monotonic walk model instead of the random walk model in standard linear filtering. The batch estimate is a solution of Quadratic Programming (QP) problem. This paper shows that the QP has a form of isotonic regression (IR) and has a linear computational complexity.The filter is implemented in a Moving Horizon Estimation (MHE) setting. The data beyond the estimation horizon are replaced by the initial condition parameters (arrival cost). The MHE for IR is nonsmooth, so the existing nonlinear MHE theory is not applicable. By exploiting properties of the IR solution, we develop an update of the MHE arrival cost, which is provably close to the full information MAP solution and stable. The analysis is complemented by a Monte Carlo simulation study of the proposed nonlinear filtering algorithm. The simulation results confirm improved performance of the proposed filter compared with a linear filter and the earlier version of the MHE update.

show abstract

“…In [22] it is shown that for common pdfs (such as Gaussian noise, Uniform noise, Laplacian noise, exponential noise) the log term in the cost function of Eq. (18) is convex in input vectors, and hence the fault diagnosis problem remains convex if G k (·) is convex (or can be made convex via bounding techniques) 10 .…”

Section: -Optimization Problem For Fault Estimationmentioning

confidence: 99%

Optimal Model-Based Fault Estimation and Correction for Particle Accelerators and Industrial Plants Using Combined Support Vector Machines and First Principles Models

Sayyarrodsari¹,

Schweiger²

2010

View full text Add to dashboard Cite

Model predictive estimation of evolving faults

Cited by 5 publications

References 10 publications

Early Detection of Solid Rocket Motor Case-Breach Failures

Early Detection of Solid Rocket Motor Case-Breach Failures

Efficient filtering using monotonic walk model

Optimal Model-Based Fault Estimation and Correction for Particle Accelerators and Industrial Plants Using Combined Support Vector Machines and First Principles Models

Contact Info

Product

Resources

About