Muhittin Yilmaz

Design-Oriented Enhanced Robotics Curriculum

This paper presents an innovative two-course, laboratory-based, and design-oriented robotics educational model. The robotics curriculum exposed senior-level undergraduate students to major robotics concepts, and enhanced the student learning experience in hybrid learning environments by incorporating the IEEE Region-5 annual robotics competition with open-ended design challenges, by establishing a robotics club, and by implementing a K-12 mentorship program. A four-faculty team developed two elective courses in which the theoretical concepts underlying the robotics design competition topics were supported with corresponding laboratory activities. Students took the courses sequentially. They were formed into diverse teams, whose performance was assessed via weekly team presentations and participation in an outreach day. The best robot design teams participated in the IEEE Region-5 competitions. All students participated in a service-learning activity, acting as mentors during local K-12 robotics competitions while enhancing their own robotics comprehension. The robotics curriculums two-year evaluation results illustrate the consistent efficacy of the curriculum during the project duration, indicating a successful robotics educational model for other academic institutions to follow.

A Smart Grid Intelligent Control Framework

This paper presents an intelligent control framework in smart grids that integrate three renewable and nonrenewable energy sources for superior economical and environmental operations. The hydro and steam power plants are considered to be controllable energy sources while the wind turbine is treated as an exogenous energy source. The proposed intelligent feedback control framework optimizes the smart grid power generation by perfectly tracking the load demand fluctuations or by maximizing economical and ecological benefits under an uncertain renewable wind energy source. Numerical results demonstrate that the proposed framework is a viable approach to integrate various energy sources for real-time smart grid implementations.

A H-infinity Control Approach for Oil Drilling Processes

Abstract This paper presents a robust optimization framework to improve the Managed Pressure Drilling (MPD) process for safer and superior economical and environmental operations while removing risk-prone conventional drilling limitations such as a need for constant monitoring of the system parameters. The nonlinear MPD process considers the mud pump flow rate and the differential flow rate of the backpressure pump and the choke as the two inputs while the process downhole pressure rate as the output. The MPD process unmodeled disturbances, uncertain geological parameters and related model nonlinearities are considered to be the corresponding system uncertainties in a closed loop robust control and optimization framework for real-time operations. Moreover, the MPD process inputs are formulated to remain within practical bounds by introducing performance weighting functions. The proposed framework numerical results demonstrate the efficiency of the closed loop robust control implementations for efficient drilling operations in operator guidance systems and provide a low-computational complexity design algorithm for safer drilling operations in regions with a-priori unknown geological properties.

A Smart Grid Robust Optimization Framework

Abstract This paper presents a robust optimization framework that integrates three energy sources for superior smart grid economical and environmental operations. The renewable hydro and wind energy sources as well as nonrenewable coal-fired steam energy source are considered to be connected via an asynchronous link. The robust control framework treats the hydro and steam power plants as controllable energy sources while it treats the wind turbine as an exogenous energy source. The proposed robust system framework ensures optimal smart grid power generation for acceptable load demand tracking or for ecological benefits under wind power and model uncertainties. Simulation results demonstrate the efficiency of the proposed framework for superior smart grid power source integration under uncertain operation conditions.“

A two-step optimization of the centro-hermitian form in direct data domain least squares approach

This paper presents a new convex solution algorithm for the direct data domain least squares (D3LS) real weight approach that is extensively used during adaptive beam pattern synthesis of smart antenna arrays. The original D3LS approach is reformulated by using centro-hermitian matrix properties as well as convex optimization techniques and solved by using the proposed two-step convex optimization framework that is implemented in linear programming framework. The first convex formulation determines the unknown coefficient that is required during the centro-hermitian matrix manipulations. The second convex formulation solves the D3LS approach to calculate the corresponding optimal real weight coefficients. The numerical beam synthesis results demonstrate that the proposed convex solution based on centro-hermitian matrix manipulations of the D3LS approach effectively solves the real weight coefficients, overcomes possible matrix inversion issues and eliminates complexity, hardware implementation and energy consumption concerns due to complex weight coefficients. Moreover, although traditional beam pattern syntheses with real weights yield symmetric beam patterns, the proposed convex solution approach successfully removes this restriction.

A Hydro Power Plant Linear Parameter Varying Control Framework

This paper presents a Linear Parameter Varying (LPV) control methodology for a hydro power plant for potentially superior smart grid implementations. The hydro power plant is assumed to be decomposable to its subsystems whose characteristics may involve different dynamical behaviors related to a real-time time-varying parameter that can be measurable in future periods. The nonlinear hydro plant dynamics are expressed in terms of a polytopic parameter-dependent model to efficiently characterize the plant dynamical changes, and the associated LPV controller synthesis perspectives are detailed. The LPV model closed loop controller synthesis and simulation results illustrate the effectiveness of the framework for nonlinear power plant optimization.

An Intelligent Control Approach for Oil Drilling Processes

Abstract The paper presents an intelligent control and optimization framework for managed pressure drilling systems. The nonlinear drilling process model was configured in a closed-loop feedback control framework to optimize the oil drilling process performance. Two main process components, namely, the mud pump flow rate and the differential flow rate of the backpressure pump and the choke, are assumed to be the control inputs while the process down hole pressure rate is treated as the process output. The control, optimization and automation of the drilling process are investigated by designing an intelligent fuzzy logic controller in a tracking problem for real-time implementation, by utilizing the closed loop system tracking error and the error rate as the controller inputs and by generating incremental changes for the two process inputs. Although the proposed control system framework is inherently nonlinear due to the nonlinear process model and the nonlinear intelligent control, the process control input and output parameters have been successfully achieved. The proposed control framework simulation results clearly illustrate that the managed pressure drilling process can be optimized as a closed loop control tracking problem, effectively removing the need for complex controller design and allowing real-time implementation in manufacturing operations in operator support systems.

Intelligent modeling and prediction of nanostructural behavior of humidity sensors

This paper presents the nanostructural characterization of a humidity sensor and obtains its intelligent models by using artificial neural networks and adaptive neuro-fuzzy inference systems. Three major processing components are manipulated via the sol-gel spin-coating technique to prepare an active element to detect humidity variations. The corresponding nanostructural analysis of the humidity sensor using X-ray powder diffraction generated the lattice distortion, known as tetragonality, data. Consequently, the three fabrication input components and the one output component of the humidity sensor data were used to obtain and validate intelligent models of the humidity sensor. The intelligent model and validation results demonstrate that the intelligent models of humidity sensors are suitable to improve the sensor fabrication process, to predict the sensor behavior under various conditions, and to help detect faulty humidity sensors.

Smart-Grid Backbone Network Real-Time Delay Reduction via Integer Programming

This research investigates an optimal delay-based virtual topology design using integer linear programming (ILP), which is applied to the current backbone networks such as smart-grid real-time communication systems. A network traffic matrix is applied and the corresponding virtual topology problem is solved using the ILP formulations that include a network delay-dependent objective function and lightpath routing, wavelength assignment, wavelength continuity, flow routing, and traffic loss constraints. The proposed optimization approach provides an efficient deterministic integration of intelligent sensing and decision making, and network learning features for superior smart grid operations by adaptively responding the time-varying network traffic data as well as operational constraints to maintain optimal virtual topologies. A representative optical backbone network has been utilized to demonstrate the proposed optimization framework whose simulation results indicate that superior smart-grid network performance can be achieved using commercial networks and integer programming.

Hydro plant network control LPV framework

This paper presents a network controlled Linear Parameter Varying (LPV) control framework for a hydro power plant. The linearized plant dynamics is assumed to be controlled via remote controller operations over communication networks. The controller action to hydro plant and the plant output to controller input signals are assumed to be transmitted over different networks with a bounded network delay and associated packet drops. As the network delay is modeled as a Pade approximation, the control system signal transmissions over the networks are modeled as an LPV system such that the accurately received packets and lost packets are used as the real-time time-varying parameter that can be measurable in future periods. The communication network LPV characteristics are expressed in terms of a polytopic parameter-dependent model to efficiently characterize the overall network operations for received and lost packages, and the associated LPV controller synthesis perspectives are detailed. The networked controlled LPV controller synthesis and simulation results clearly demonstrate the effectiveness of the framework for networked control system delay and packet loss issues on stability and performance.