Research Associate in Computational Neuroscienec
Academic Division: Information Engineering
Research group: Computational and Biological Learning
Telephone: +44 1223 7 48520
Mahdieh's research interests are generally centered around robust fixed-structure control of large-scale uncertain systems with several applications: microgrids, (smart) power grids, and computational neuroscience.
Research at EPFL. The work during Mahdieh’s PhD studies at EPFL was mainly concerned with robust fixed-structure control of linear time-invariant (LTI) systems affected by uncertainty. She developed new solutions in terms of a convex optimization problem for fixed-structure control of polytopic systems. She also addressed the problem of control configuration design in large-scale interconnected systems with polytopic uncertainty. Her results show that the control configuration design is achieved by solving a convex optimization problem whose solution delivers a trade-off curve that starts with a centralized controller and ends with a decentralized or a distributed controller. She visited LAAS-CNRS in Toulouse, France in April 2015 to start collaboration with Dr. Dimitri Peaucelle. The collaboration has led to a recent survey paper on static output feedback in http://www.sciencedirect.com/science/article/pii/S1367578816300852.
The application part of her dissertation focused on control challenges associated with the integration of renewable energy resources, such as wind and photovoltaic, into power grids. The main motivation behind her studies was the European Union’s aggressive environmental targets which should be met by 2020. The targets are: 20% reduction of greenhouse gas emissions, 20% of energy generation from renewable sources and a 20% reduction in primary energy use due to efficiency improvements. Her research specifically dedicated to voltage control of microrgids which are low-voltage small electrical networks heterogeneously composed of distributed generations (DGs), loads, and energy storage systems connected to the main grid at a single point of connection, the Point of Common Coupling (PCC). One of the main issues in the control of microgrids is plug-and-play (PnP) functionality of DGs and microgrid topology change. DGs frequently join and leave the power generation system due to availability and intermittency of renewable energies, an increase in energy demand, faults, maintenance, etc. Under PnP operation, different DGs are arbitrarily plugged-in or plugged-out from the microgrid; however, voltage and frequency of the local loads have to be stabilized without retuning the microgrid control system, in the absence of any communication link. Therefore, a decentralized control strategy is necessary to guarantee the stability of the microgrid system in the case of PnP functionality of DGs. In her PhD thesis, a solution for the problem of PnP functionality of DGs was presented. She showed that an inverter-interfaced microgrid consisting of multi DGs under plug-and-play functionality can be cast as an LTI dynamical system subject to polytopic uncertainty. By virtue of this novel description and use of the results from theory of robust control, the stability of the microgrid system under PnP operation of DGs is preserved. It was also shown that the interaction terms in the presented model of microgrids are neutral if several structural constraints are imposed on Lyapunov matrices satisfying the stability of each DG. Under those conditions, the control design procedure is scalable meaning that the synthesis of a local voltage controller uses only information of the corresponding DG and transmission lines connected to it.
Research at the Linköping University. Mahdieh's research as a postdoctoral fellow in the Division of Automatic Control at the Linköping University in Sweden focuses on modeling of large-scale power grids with uncertainty. She has developed a new method for modeling and analysis of large-scale nonlinear systems. Power grids are generally nonlinear complex systems. Moreover, they are usually subject to several uncertainty sources mainly including load variations (leading to different operating points). The uncertainty affects stability and deteriorates the closed-loop system performance. In order to tackle the uncertainty issue and handle system nonlinearities, a linear parameter varying (LPV) model, in which operating points are considered as varying parameters, is adopted. The main motivation for the development of LPV models is that this framework captures a large class of nonlinear systems, while ensuring computational tractability in analysis and control. Moreover, it enables us to use the linear systems theory for stability analysis and control of a wide class of nonlinear systems. To this end, nonlinear equations in power grids mostly associated with synchronous generators are approximated with an LPV model based on a family of linearized models around different operating points.
Current Research at the University of Cambridge. Her challenging research at CBL is at the intersection of control engineering and computational neuroscience. The main idea is to use the tools from control theory, more specifically fixed-structure control, for dynamics of computation in brain circuits. The research focuses on a randomly connected recurrent network of firing-rate units modeled as an interconnected LTI system with a connection matrix. The connection matrix in the network is designed in order to stabilize a state in which all neurons fired at low and constant firing rates. However, the connection matrix is under some structural constraints since it obeys Dale’s principle which states each neuron only excites or only inhibits, but not both. Due to the Dale’s law, the connection matrix is composed of separate positive and negative columns. Therefore, the problem can be formulated as a fixed-structure control problem which lies at Mahdieh’s area of expertise.
I am a control scientist, but working with neuroscientists in the Computational and Biological Learning (CBL) group at the University of Cambridge. I am supervised by Dr. Guillaume Hennequin. Our project lies at the intersection of control theory and computational neuroscience.
Research Associate, Computational and Biological Learning Group, Department of Engineering, University of Cambridge, UK (February 2017 - now).
Postdoctoral Fellow, Division of Automatic Control, Electrical Engineering Dep., Linköping University, Linköping, Sweden (May 2016 - January 2017).
Research Assistant at Automatic Control Laboratory, Ecole Polytechnique Federale de Lausanne (EPFL), Switzerland (January 2011 - February 2016).
Visiting Scholar at LAAS-CNRS, Toulouse, France (April 2015).
Visiting Scholar at the Département de Génie Électrique, École Polytechnique de Montréal, Montréal, Canada (April 2014 - July 2014).
Mahdieh is currently a research associate in the Computational and Biological Learning (CBL) Group at the Department of Engineering, the University of Cambridge, UK. Prior to that, she was a postdoctoral fellow in the Division of Automatic Control at the Department of Electrical Engineering, Linköping University in Sweden. She received her Ph.D. in Systems and Control Theory from Electrical Engineering Department, Swiss Federal Institute of Technology in Lausanne (EPFL), Switzerland in February 2016. She obtained a BSc and MSc with honors in Electrical Engineering from Tehran Polytechnic. She was a visiting scholar at LAAS-CNRS in Toulouse, France and Ecole Polytechnique de Montreal in Montreal, Canada. Her research interests are centered around fixed-structure controller design, control of large-scale uncertain systems, and their applications to energy systems, microgrids/smart power grids, and computational neuroscience.