Efficient battery system simulations with the Discontinuous Galerkin method applied to the SPMe battery cell model
2024 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE credits
Student thesisAlternative title
Effektiva simuleringar av batterisystem med den Diskontinuerliga Galerkin-metoden applicerad på battericellsmodellen SPMe (Swedish)
Abstract [en]
Throughout the past few decades, significant steps have been made towards efficient and accurate simulation of battery cells. Just a few years ago, it was considered state-of-the art to simulate a single battery cell in real-time. These days such simulations are much faster than real-time due to development of new methods and general increase in computational power.
While these advances are remarkable, it can still be a significant challenge to go from simulating a single cell to thousands of coupled cells as when simulating a battery system. Such simulations are treated by this thesis, which is done in two ways.
First, the Discontinuous Galerkin (DG) method is applied to an electrochemical battery cell model - the Single Particle Model with Electrolyte (SPMe). The application of this method is validated against both analytic solutions and reference simulations. Convergence rates and performance is studied with respect to adaptive time integration, and recommended discretization parameters for the problem is given.
Second, the electrical coupling of the battery cells within a battery system is treated with graph-theoretical methods. After first translating the electrical circuit into a graph, a minimum cycle basis is used to derive sufficient equations to resolve Kirchhoff's laws. This yields a large system of equations, which is reduced to a much smaller size by a reduction technique.
It is then shown that the performance of the methods can achieve faster than real-time simulations of battery systems with thousands of cells, highlighting the potential of the methods of this thesis.
Abstract [sv]
Under de senaste decennierna har betydande framsteg gjorts mot effektiv och exakt simulering av battericeller. Bara för några år sedan ansågs det vara ett betydande resultat att simulera en enstaka battericell i realtid. Idag är sådana simuleringar mycket snabbare än realtid tack vare nya metoder och ökad beräkningskraft.
Trots att dessa framsteg är enastående så kan det vara en betydande utmaning att gå från att simulera en enskild cell till tusentals kopplade celler, som vid simulering av ett batterisystem. Sådana simuleringar behandlas i detta examensarbete, vilket görs på två sätt.
Först appliceras den Diskontinuerliga Galerkin-metoden (DG) på en elektrokemisk battericellsmodell - Single Particle Model with Electrolyte (SPMe). Metoden valideras emot både analytiska lösningar samt referenslösningar. Sedermera studeras konvergenstakt och prestanda, i synnerhet med avseende på adaptiv tidsintegrering. Rekommenderade diskretiseringsparametrar för problemet ges därefter.
Därefter behandlas den elektriska kopplingen av battericellerna inom ett batterisystem med grafteoretiska metoder. Efter att först ha överfört den elektriska kretsen till en graf appliceras en minimal cykelbas för att härleda ekvationer som uppfyller Kirchhoffs lagar. Detta resulterar i ett stort ekvationssystem, som reduceras till en mycket mindre storlek med en reduceringsteknik.
Det visas till sist att prestandan hos metoderna uppnår hastigheter snabbare än realtid för simuleringar av batterisystem med tusentals celler, vilket visar på möjligheterna med metoderna i detta examensarbete.
Place, publisher, year, edition, pages
2024. , p. 125
Series
TRITA-SCI-GRU ; 2024:424
Keywords [en]
battery simulation, battery systems, battery cells, Discontinuous Galerkin (DG) method, graph theory, Single Particle Model with Electrolyte (SPMe), Minimum cycle basis, electrical circuits
Keywords [sv]
batterisimulering, batterisystem, battericeller, Diskontinuerliga Galerkin-metoden (DG), grafteori, Single Particle Model with Electrolyte (SPMe), minimal cykelbas, elektriska kretsar
National Category
Computational Mathematics
Identifiers
URN: urn:nbn:se:kth:diva-359218OAI: oai:DiVA.org:kth-359218DiVA, id: diva2:1932349
External cooperation
Northvolt Systems AB
Educational program
Master of Science - Applied and Computational Mathematics
Supervisors
Examiners
2025-01-292025-01-292025-01-29Bibliographically approved