Ionel, Dan

Summary of Research Activities

SPARK Lab
 

The SPARK conducts academic research, some of which is sponsored by government and industry organizations.  Main areas of work include future generation of electric machines and drives, electric aircrafts and vehicles, wide band gap device based power electronics converters, and utility and distribution scale energy storage.

Future Generation Electric Machines and Drives —The machines and drives research in the SPARK lab builds upon the latest advancements in power electronics, materials and finite element computation technologies to develop transformational highly torque dense, efficient electric machines.  Large scale stochastic optimization techniques such as differential evolution are being employed to select the best designs meeting application requirements. Examples of specific research topics include electric machine topologies with advanced materials, the development of optimization algorithms, wide band gap semiconductor drives, and detailed system level modeling. Sponsors, collaborators and supporters include NSF, NASA, ANSYS Inc., and Regal Beloit Corporation.

Power Electronic Systems with Wide Band Gap Devices— Power electronics converters are critical components of solar energy conversion, grid power conditioning and motor drive systems. The DOE’s Solar Technology Office’s recent Sunshot Initiative has the goal of enabling solar energy’s costs to be competitive to that conventional energy. The recent availability of wide band gap (WBG) devices creates many opportunities for innovative power electronics designs which improve efficiency, reliability and power density of motor drives and solar converters. Examples of specific research topics in the lab include – high performance solar inverters, and multilevel converters. Sponsors, collaborators and supporters include DOE and NASA.

Electric Vehicles and Aircraft—The SPARK lab has been supporting the design and electromagnetic analysis of the electric motor used in the UK Solar Car. In this regard, analysis was conducted on advanced motor topologies, and their control.  In addition, the SPARK lab is also developing collaborations with NASA on the design of the next generation electric aircraft, which employ electric machines in varying degrees for propulsion.  The electricity required may be stored in batteries, or be generated on board from turbo-generators, or using solar PV panels.  The SPARK lab’s specific areas of work within this topic include power system architectures, sizing of engines, motors, and energy storage, and power electronics converter topologies. Sponsors, collaborators and supporters include NASA.

Utility Scale and Distributed Energy Storage for Future Power System and Smart Grid — The electric grid is changing owing to the increasing penetration of renewable energy. The intermittent availability motivates their use in conjunction with energy storage systems such as batteries, which are becoming popular owing to their modularity and easier controllability. Specific research topics in the SPARK lab include – utility scale energy storage, and home energy storage and demand response and developing collaboration activities involve LG&E/KU, EPRI,  and A.O Smith Corp., and the DOE.

Students

Peng Han, PostDoc

Vandana Rallabandi, PostDoc

Bryan B Poe, Staff

Christopher Heintz, Grad

Narges Taran, Grad

Damien Lawhorn, Grad

Yibin Zhang, Grad

Huangjie Gong, Grad

Oluwaseun Akeyo, Grad

William Song, UGrad

Monon Rahman, UGrad

Luke Wormald, UGrad

Rayna A Weibrecht, UGrad

Thomas Barber, UGrad

Kathryn G Schantz, UGrad

Murat Gurhan Kesgin, UGrad

Rosemary E Alden, Non-work study

Xin Zhang

Daniel Lemon

Yaser Chulaee, Added 03/08/2021

Damien L Lawhorn, Grad, Added 03/08/2021

Projects on the HPC System

Large Scale Optimization of Electric Machinery—The SPARK Lab has a very large number of licenses to ANSYS/Maxwell as a result of a collaborative project with ANSYS. The projects that benefit by them include an NSF, and a Regal Beloit sponsored project. These involve the optimization of axial flux electric machines, which tend to be more compact and efficient than their conventional radial flux counterparts, but cannot be analyzed by the widely used two-dimensional techniques because of their complex geometry, and necessitate the use of computationally expensive three dimensional finite element analysis.  The performance parameters which need to be evaluated include the losses, which necessitates time transient analysis with motion. To ensure accuracy of the calculations, several time steps in one electrical cycle need to be computed, and problems involving eddy currents especially require a large number of mesh elements, of the order of millions. Solution of a single problem of this scale would take inconceivably long on a conventional computing system, indeed, only the problem discretization or meshing, takes days.

 It is therefore expected that running an optimization algorithm, which requires the evaluation of a thousand candidate designs, would take a thousand times longer, in other words, it would be virtually impossible to solve on the standard desktops within the SPARK lab.   ANSYS/Maxwell allows for job distribution across multiple cores, and can leverage the HPC cluster. This would reduce the computation time substantially, making it possible to employ optimization algorithms for axial flux machines. Therefore, this project would directly benefit from the HPC system.

UK Personnel: Narges Taran (Student/RA), Murat Kesgin (Student) Vandana Rallabandi (Post doc) and Dan M. Ionel (Faculty)

Non – UK: Greg Heins, Dean Patterson, Mark Theile (Regal Beloit Corporation)

Complex Power System Studies on Electric Aircraft—This project involves the design and optimization of drive train architecture, and power electronics converters for electric airplanes. The aircraft power system is a complex one, and requires extensive simulation studies, at different time steps, to accommodate fast electrical as well as slow mechanical transients.  Different components of an aircraft system include the jet engine, electric machine, power electronics converter, and battery. One of the objectives is to optimize for efficiency and weight, and different scenarios of size and number of engines, motors, and batteries need to be considered. The platform for simulations is ANSYS/Simplorer, which allows for job distribution across multiple cores, and can directly leverage the HPC system.

UK Personnel: Damien Lawhorn (Student/RA) and Dan M. Ionel (Faculty)

Non – UK: Ralph Jensen (NASA Glenn Research Center)

Large-scale Simulation of Smart Buildings and Power Electronics Controls—The SPARK Lab, works with various industrial partners and utility companies on smart net zero energy buildings. Net zero energy buildings aim to feed back to the grid as much energy as they absorb, and often include rooftop solar panels. Issues such as overloading of distribution lines, transformers, and stability issues arise because the peak because the peak building loads are seldom coincident with the maximum solar generation. The project involves scheduling of the battery energy storage in these buildings to achieve different objectives including peak reduction, and load unbalance reduction. These tasks involve simulation of the buildings to calculate their energy consumption, 

power system network solution to compute the power flow at each node. The building models are dynamic, and require time stepped solutions. The buildings include power electronics systems, which also require high resolution time transient solutions.

The problem is large scale, as each network can contain up to hundreds of buildings. Moreover, the optimization of the operation of the battery energy storage system would necessitate the simulation of thousands of schedules. If this problem were to be solved using standard desktop facilities within the SPARK lab, the solution time may prove to be prohibitively high. With the HPC system, several building simulations can be executed in parallel, drastically reducing the solution time. The software framework for this project is being developed within SPARK and includes Python, EnergyPlus and OpenDSS.  These software tools are open source.  The power electronics simulations will be carried out using ANSYS/Simplorer.

UK Personnel: Huangjie Gong (Student/RA), Akeyo Oluwaseun (Student/TA), Yibin Zhang (Student/TA), Vandana Rallabandi (Post doc) and Dan M. Ionel (Faculty).

Software Framework and Computational Methods

The computational methods and software being used on these projects are MATLAB and ANSYS/Maxwell and ANSYS/Simplorer. The SPARK Lab has an official agreement with ANSYS which includes a very large number of licenses.  The software framework for the smart buildings project which uses Python, EnergyPlus and OPenDSS is being developed.