Page 75 - ASME DSCC 2015 Program

Technical Program

An Adaptive Robust Control for Hard Rock Tunnel Boring Machine
ConTRIBuTED SESSIon
1-5-1 fM3 uncertain Systems and Robustness Cutterhead Driving System
George Bellows E 1:30pm–3:30pm Contributed regular paper. DSCC2015-9697
chengjun Shao, jianfeng Liao, xiuliang Li, hongye Su, Zhejiang
Session Chair: Rama Yedavalli, Ohio State University University, Hangzhou, China
Session Co-Chair: Roger fales, University of Missouri
The cutterhead driving system of TBM is one of the key components for
rock cutting and excavation. In this paper, a generalized nonlinear time-vary-
Stabilization of Inverted Pendulum on a Cart in the Presence of
ing dynamic model is established for the hard rock TBM cutterhead driving
uncertainties
system. Parametric uncertainties and nonlinearities and unknown distur-
Contributed regular paper. DSCC2015-9975
bances exist in the dynamic model. An adaptive robust control strategy is
Joonho lee, Jongeun Choi, Michigan State University, East Lansing, MI, proposed to compensate the uncertainties and nonlinearities to achieve
United States
precise cutterhead rotation speed control. In order to simulate the compre-
This paper presents an output feedback control design to stabilize the hensive performances of ARC controller three different kinds of external
inverted pendulum at the upright equilibrium as an extension of our previous force disturbances are added in this model. Compared to the traditional PID,
work [1]. Compared to our previous work, we add one more time scale be- ARC can effectively handle the different kinds of external force disturbances
tween a pendulum angle and angular velocity to reduce a traveled distance with sufficient small tracking errors.
of the cart. State feedback control is designed to enable the pendulum to
ConTRIBuTED SESSIon
pass through input singularity configurations. Extended High-Gain Observ-
1-27-2 fM5 Modelling and Validation 2
ers are used to estimate velocity and acceleration terms while dynamic Emerson Burkhart A 1:30pm–3:30pm
inversion utilizes the estimates to deal with input coefficient uncertainties
and singularity configurations. The proposed control is verified through
Session Chair: David Bevly, Auburn University
numerical simulations.
Session Co-Chair: Christopher Pannier, University of Michigan, Ann Arbor
a New, Necessary and Sufficient Vertex Solution for robust Stability
Check of unstructured Convex Combination Matrix families Comparative Evaluation of Control-oriented Zone Temperature
Contributed regular paper. DSCC2015-9986 Prediction Modeling Strategies in Buildings
Rama Yedavalli, Ohio State University, Columbus, OH, United States Contributed regular paper. DSCC2015-9864
A New, Necessary and Sufficient Vertex Solution for Robust Stability Check Venkatesh chinde, jeffrey heylmun, adam kohl, Zhanhong jiang,
Soumik Sarkar, Atul Kelkar, Iowa State University, Ames, IA, United States
of Unstructured Convex Combination Matrix Families
Predictive modeling of zone environment plays a critical role in developing
Parameterized uncertainty Model using a Genetic Algorithm with
and deploying advanced performance monitoring and control strategies for
Application to an Electro-Hydraulic Valve Control System
energy usage minimization in buildings while maintaining occupant comfort.
Contributed regular paper. DSCC2015-9994
The task remains extremely challenging, as buildings are fundamentally
Zuheng Kang, Bahaa Kazem, Roger fales, University of Missouri, MO, complex systems with large uncertainties stemming from weather, occu-
United States
pants, and building dynamics. Over the past few years, purely data-driven
This work proposes a new method of determining a parameterization of an various control-oriented modeling techniques have been proposed to
uncertainty model using a genetic algorithm. A genetic algorithm is used address different requirements, such as prediction accuracy, flexibility,
in a unique way to solve the non-convex parameterization problem in this computation and memory complexity. In this context, this paper presents a
work. The methods presented here are demonstrated on an electrohydrau- comparative evaluation among representative methods of different classes
lic valve control system problem. This demonstration includes parameteriz- of models, such as first principles driven (e.g., lumped parameter autoregres-
ing an uncertainty class determined from test data for 30 replications of an sive models using simple physical relationships), data-driven (e.g., artificial
electrohydraulic flow control valve. The parameterization of the uncertainty neural networks, Gaussian processes) and hybrid (e.g., semi-parametric).
is used to analyze the robust stability of a control system for a class of Apart from quantitative metrics described above, various qualitative aspects
valves. such as cost of commissioning, robustness and adaptability are discussed
as well. Real data from Iowa Energy Center’s Energy Resource Station (ERS)
test bed is used as the basis of evaluation presented here.

75