- Profile Control of
The spatial distribution of
system properties offers additional
challenges for control of the same. A common strategy is explicit
control of the endpoint property while ignoring the history of
processing along the reaction path. We are exploring the idea of
exploring the control of the entire property profile along the spatial
direction. It is expected that such an approach will enable a tight
regulation of history-dependent properties (such as fiber length in a
pulp digester, particle size distribution in polymerization reactor).
Further, upstream disturbances may be rejected prior to their effects
being felt on the endpoint property. Currently, we are working on
observer design for a pulp digester for building inferential
measurements of the property at various locations of the digester.
These soft measurements will consitute the property profile, which will
be controlled using NMPC.
- Optimization and Grade
Transition Control (Collaborator: Prof. R.D. Gudi, IIT Bombay)
Grade transition is a
frequent operation in the process industry
(polymer, pulp and paper). Optimal grade transition recipes typically
target minimization of transition time and production of
off-specification material. Generation of the optimal recipes requires
solution to a dynamic optimization problem, which has been approached
using gradient-based methods such as SQP (Wang et al., 2000) and
differential evolution (DE) methods (Mandal et al., 2003). While
gradient-based methods are well known for computational efficiency,
they typically fail to provide global optimum (unless the convexity
assumption is invoked). Further, they cannot handle discontinuous
derivatives of objective functions and constraints. On the other hand,
DE can potentially provide the global optimum and is not susceptible to
non-smoothness of the objective function and constraints. However, DE
methods require numerous function evaluations thereby making the
algorithm computationally expensive. We are currently working on a
hybrid approach that can combine the global optimum properties of DE
and faster convergence rates of SQP.
- Systems Biology
Prof. K.V. Venkatesh, IIT Bombay)
Living systems must adapt
quickly and stably to uncertain environments.
A common theme in cellular regulation is presence of multiple feedback
loops in the network. An example of such a feedback structure is
regulation of tryptophan concentration in Escherichia coli. A pertinent
question is whether such multiple feedback loops is a case of
regulatory overkill, or do these different feedback regulators have
distinct functions? (Freeman, Nature, 295, 313-319, 2000). Another
moot question is how can robustness to uncertainties be achieved
structurally through biological interactions. We are currently
establishing correlations between the feedback structure and robustness
using feedback theory. We are also focussing on how feedback designs in
nature cope with intrinsic noise.
- N.U. Padhiyar, and S. Bhartiya, “Profile Control in Distributed Parameter Systems Using Lexicographic Optimization Based MPC”, Journal of Process Control (In Press) .
- N. Nandola and S. Bhartiya , “A multi-model framework for control of hybrid systems”, Journal of Process Control, 18, 131-148 (2008).
- N. U. Padhiyar, A. Gupta, A. Gautam, S. Bhartiya, F.J. Doyle III, S. Gaikwad and S. Dash, “Nonlinear Inferential Multi-Rate Control Of Kappa Number At Multiple Locations In A Continuous Pulp Digester” Journal of Process Control, 16, 1037-1053 (2006).
- S. Bhartiya, N. Chaudhary, K.V. Venkatesh, and F.J. Doyle, “Multiple feedback loop design in the tryptophan regulatory network of Escherichia coli suggests a paradigm for robust regulation of processes in series”, Journal of the Royal Society Interface, 3, 383-391, (2006)
- A. Ruhela, M. Verma, J.S. Edwards, P.J. Bhat, S. Bhartiya, K.V. Venkatesh, “Autoregulation of regulatory proteins is key for dynamic operation of GAL switch in Saccharomyces cerevisiae”, FEBS Letters, 576, 119-126, (2004)
- K.V. Venkatesh, S. Bhartiya, and A. Ruhela, “Multiple feedback loops are key to a robust dynamic performance of tryptophan regulation in Escherichia coli”, FEBS Letters, 563, 234-240, (2004).
- S. Bhartiya, P. Dufour and F.J. Doyle III, "Fundamental thermal-hydraulic continuous pulp digester model with grade transition", AIChE Journal, 49, 411-425 (2003).
- S. Bhartiya and J.R. Whiteley, "Factorized approach to nonlinear MPC using a radial basis function model", AIChE Journal, 47, 358-368 (2001).