Bath PhD Studentship Competition: Physics-Based Battery Models for Next-Generation Battery Management

Bath PhD Studentship Competition: Physics-Based Battery Models for Next-Generation Battery Management

Quick Decision

This Bath Mechanical Engineering PhD targets a specific gap between electrochemical insight and a working battery-management system: retaining useful DFN/P2D physics within a microcontroller's time and memory limits. It links analytical transport, adaptive model fidelity, degradation inference and embedded C/C++ validation. The distinctive opportunity is to study when a cheaper model remains physically adequate, rather than treating voltage-fit accuracy or desktop simulation speed as sufficient evidence of deployability.

The opportunity is Strong but verify fit, with a substantial mathematical and implementation expectation. The announcement says the successful candidate should have demonstrable experience in PDE modelling, electrochemical modelling, parameter identification, embedded implementation and high-performance computing. These are not all optional interests. A master's qualification is advantageous. Home and Overseas candidates can compete for URSA tuition and maintenance support; personal selection, the 2027 stipend and paid months are not confirmed. The project deadline is 29 November 2026; current faculty URSA policy adds 23:59 GMT. Expected entry is 4 October 2027, conditional on funding, with alternatives by supervisor/Doctoral College agreement. A useful first contact with Hao Yuan, hy310@bath.ac.uk, should address the model/platform/data baseline and how existing experience meets the stated expectation.

Research Problem and Doctoral Work

Physical fidelity has a computational price

The Doyle–Fuller–Newman model, also called pseudo-two-dimensional or P2D, couples transport through the electrodes/electrolyte with diffusion inside representative active-material particles and electrochemical/electrical behaviour. Its richer state description can become expensive for an embedded BMS, especially under high-rate, rapidly changing conditions. The research question is how much of that description can be preserved at an acceptable computational cost without concealing important limitations.

The advertised approach builds on analytical and semi-analytical solid-phase diffusion in spherical particles, combined with efficient electrolyte transport. A closed-form solution to a simplified subproblem does not make the entire nonlinear, coupled DFN system exact. The relevant research contribution lies in the assumptions, numerical conservation and error introduced when those components are combined, as well as the implementation cost.

The proposed hierarchy connects efficient reduced models, including SPMe, with higher-fidelity P2D formulations. Selection is intended to use physically motivated validity indicators—electrolyte gradients, reaction limitations, overpotential and dimensionless transport/kinetic criteria—rather than fixed operating thresholds. This is an advertised research objective. The switching logic, accuracy envelope and evidence that switching preserves consistent states have not yet been supplied.

Inference, degradation and embedded validation

The project also proposes descriptions of SEI growth, lithium plating, mechanical cracking and electrolyte depletion, supporting physics-informed SOH, capacity and resistance-fade estimation. These mechanisms should not be treated as interchangeable labels. Several parameter combinations or degradation pathways may produce similar terminal signals; a good voltage fit alone does not establish a unique internal mechanism. Distinguishing structural identifiability, practical information content and uncertainty is therefore central to judging the research, not merely an optional optimisation improvement.

Models will be implemented in embedded C/C++ and assessed for accuracy, stability, memory and execution time. Validation is planned against higher-fidelity P2D/DFN simulations and commercial-cell experiments under high-rate/dynamic use, including UDDS and WLTP, with temperature and ageing considered. Faster execution, lower memory and improved stability remain targets, not achieved specifications or guarantees of BMS qualification. No processor, cell chemistry, execution deadline or numerical tolerance is advertised.

An analytical preparation framework would separate numerical verification against a reference solver from validation against independent cell observations. Reference-model agreement can expose implementation errors but cannot prove that both models describe the real cell. Similarly, an average desktop runtime does not determine worst-case embedded latency, and a successful drive-cycle fit may not transfer to different temperature or ageing conditions. Memory, precision, state consistency and estimator robustness need to be assessed on the actual target. These are decision questions, not a prescribed controller or cell-testing procedure.

A manageable thesis would need an agreed priority among transport approximation, adaptive fidelity, degradation and parameter inference. Attempting every mechanism and every chemistry simultaneously could dilute the validation. The announcement supplies a clear direction but not that prioritisation, experimental allocation, data-ownership terms or the balance between offline HPC work and deployed embedded computation.

School, Programme and Field Context

The formal route is the University of Bath Mechanical Engineering PhD, within the Department of Mechanical Engineering and Faculty of Engineering & Design, at Claverton Down, United Kingdom. Bath awards the degree. URSA is a doctoral studentship rather than an employment contract. IAAPS is a relevant research environment, not evidence of a separate degree or joint award; participation in industrial research does not establish this student's employer or commercial secondment.

The programme covers automotive, energy systems, manufacturing/design and materials/structures. It lists IAAPS among its research environments, provides doctoral skills workshops and research supervision, and expects a substantial thesis and viva. Students normally begin as probationers and confirm registration through written work and an oral assessment. The normal full-time period is four years, with a minimum of two and a normal maximum of four; September/October and agreed mid-year starts are possible. These rules concern registration and completion, not personal funded months.

Bath is 125th overall in QS World University Rankings 2027, compared with 132nd in 2026. Its reported 92% world-leading or internationally excellent research in REF 2021 is institution-wide, not a battery-management or Mechanical Engineering group score. A current comparable BMS-specific ranking, group completion rate or project graduate-employment record was not established. The more direct subject signal is the mechanical/automotive environment and the supervisor's identified numerical-battery work.

IAAPS describes physical and virtual powertrain testing, battery-related facilities, hardware-in-the-loop capability, digital engineering and HPC at the Bristol & Bath Science Park. This makes simulation-to-system validation plausible in the wider ecosystem, but it does not reserve a cell cycler, microcontroller board, processor-in-the-loop facility, HPC allocation or industrial dataset for this PhD. Location and access should also be clarified rather than assuming all work occurs on the Claverton campus. Specialist electrochemical, numerical and embedded training, co-supervision and a personal milestone plan remain to be agreed.

Supervisor and Lab Assessment

Identity and research continuity

Hao Yuan is a Bath Lecturer in Mechanical Engineering, affiliated with IAAPS and accepting doctoral students; his ORCID is 0000-0002-6199-6142. His institutional profile displays H.Yuan@bath.ac.uk, while the project gives hy310@bath.ac.uk. His interests include numerical modelling, lithium-ion batteries and chemical kinetics. These identifiers distinguish him from other researchers named Hao Yuan at unrelated institutions. His wider combustion research is not automatically evidence of battery-model performance.

The profile lists participation in the JLR Prosperity Partnership through January 2027. That record supports an industrial research connection, not an industrial sponsor, dataset commitment or funding allocation for a student expected to enter in October 2027. No additional supervisor is formally named in this announcement. Paper co-authors and institutional collaborators should not be treated as appointed co-supervisors.

Recent originals and their limits

Zhao et al.: reduced multiphysics under charging

Zhao and colleagues, 2026, including Yuan, developed an enhanced single-particle model coupling electrochemistry, mechanics and temperature. The study uses a commercial 40 Ah NCM622/graphite cell and held-out charging rates, reporting voltage RMSE 23.43–31.46 mV. It is directly relevant to reduced-order multiphysics and multi-observable validation, but it does not establish the new adaptive P2D/C++ microcontroller framework. The available evidence comprises the complete institutional author abstract and substantive publisher-indexed model/contribution passages; the main results/discussion and supplementary files are outside that evidence range. It was online 16 July 2026; its 1 November 2026 issue date does not make it an unpublished future proposal.

Benoit et al.: sparse-label SOH estimation

The 2026 paper by Benoit, Asef, Yuan and Jafari combines Encoder/GRU features and Gaussian-process regression for semi-supervised SOH estimation using NASA dynamic-discharge data. It is relevant to sparse labels and uncertainty-aware estimation, not proof of mechanistic parameter identifiability or embedded DFN deployment. The final institutional abstract gives RMSPE 2.029%, whereas the accepted-manuscript abstract gives 2.112%; these versions must not be blended. The abstract's 15% labelled/75% unlabelled wording leaves the remaining allocation unresolved here. The evidence is limited to complete original abstracts and metadata; the main evaluation, leakage controls, code and independent replication remain unassessed.

External extended-SPM analytical solution

The 2026 Journal of Computational Science study A novel exact solution of extended single particle model… is an external original, not Yuan's paper. Primary indexed sections describe volume averaging and polynomial reduction followed by analytical solutions, electrolyte approximations and comparison with P2D/experimental NMC data. Its reported agreement concerns a particular reduced model; “exact” does not mean an exact general DFN solution or verified microcontroller speed. The evidence includes the abstract, model/solution and experimental sections, results opening and selected conclusion, but not the complete results/discussion. The primary author list was not independently confirmed, so no author attribution is inferred.

Yuan's 2021 conserved-spherical-diffusion/Jacobian–Newton–Krylov paper is useful foundational evidence but was published 10 October 2021, just outside five years by 18 October 2026. Its abstract's speed and conservation comparisons are tied to that implementation and comparator, not this new PhD. It is not counted as a third recent team original. The three 2026 studies above therefore comprise two team papers and one external comparison with explicitly limited reading depth, not three complete team-paper readings. External CPG-SPMT software also provides a transparent Python model/validation example; the evidence covers its README, without code execution or independent verification of the reported benchmarks.

Funding, Application and Living Cost

International competition and formal submission

The 2027 Engineering & Design URSA policy described in October 2026 provides for Home and Overseas tuition support, a maintenance stipend and up to £1,000 annual research-training support. The faculty's general support is up to 3.5 years, not confirmation of this individual's 42 paid months. The advertisement's £21,805 is explicitly the 2026/27 reference stipend. The 2027 personal rate, exact duration, indexation and support beyond the award remain unknown. Overseas studentships are capped, with no project-specific proportion or admission probability given. Training support is not extra living cash.

The project gives 29 November 2026 without a clock; the current faculty funding policy specifies 23:59 UK time/GMT. Expected funded entry is 4 October 2027, with alternatives by agreement. Formal application must be for Mechanical Engineering PhD, selecting University of Bath URSA under Funding Your Studies. Choose at most two projects: enter the first title and lead in the project-title and Name of intended supervisor at University of Bath fields; the second uses Project 2 Title and Name of intended supervisor for 2nd choice project at University of Bath. Incorrect entries can prevent automated consideration.

Programme and application guidance require qualifications/transcripts and translations as appropriate, CV, personal statement, English evidence and two references, including an academic reference from the most recent study. Identify at least one, up to two intended lead supervisors. The 1,000-word proposal rule concerns a self-proposed idea, not automatically this defined topic. Confirm additional project-specific material rather than inventing a mandatory 500-word summary. General international/home course deadlines do not extend the URSA deadline. Bath allows AI support for research, planning and proofreading but expects applicants to author application content.

The University encourages applications from under-represented groups. Applicants may explain circumstances that affected their educational attainment on the application form, preferably in a short paragraph at the end of the personal statement. If applicants disclose a disability in the application, Bath's Disability Service will contact them to discuss support needs.

China academic route, English and ATAS

There is a positive conditional route for mainland-China-qualified applicants: the Mechanical programme explicitly welcomes worldwide applicants with assessed equivalent qualifications, URSA includes Overseas tuition, and official international admission and Student-visa procedures exist. This is not simply an inference from a worldwide label. The project requires a first or upper-second-class honours equivalent; a master's is advantageous. The generic programme offers a relevant Master's at Merit/equivalent as an alternative academic route, subject to the project's further expectations and individual assessment.

The indexed China research-degree guidance mentions a first-class bachelor's background with research experience or a research-containing master's as desirable; the direct country page represented in the October 2026 evidence directs doctoral applicants to courses instead. The 62–75% table is taught-postgraduate guidance, not a universal doctoral cutoff. Neither compulsory master's study for every Chinese applicant nor automatic bachelor's direct entry is established. Admissions must assess the actual institution, degree, transcript and research evidence.

Mechanical requires IELTS 6.5 overall/6.0 each. TOEFL before 21 January 2026 requires 90 overall/21 each; tests on or after 21 January 2026 use 4.5 overall/4.0 each. Scores normally need to fall within 24 months of starting. English-medium-degree exemptions are individually assessed under recency/residence conditions; a Chinese medium-of-instruction letter is not an automatic waiver. Non-UK applicants not currently studying in the UK must meet the English requirement before the project deadline, with no exceptions. Any exemption needs approval in time; a general pre-sessional option is not an assurance of late funding eligibility.

Unlike an unclassified project, the Mechanical programme explicitly states ATAS coverage. Chinese nationality is not exempt under the current rules, so resolve the offer's subject/CAH and clearance requirements with admissions. This is an individual approval process, not a categorical nationality exclusion. A Student visa normally needs the appropriate offer, CAS and funds/English evidence. International students must pay visa and IHS from personal funds; current reference charges are £558 for the visa and £776 per year for student IHS. Upfront IHS depends on visa validity and part-year rules, and 2027 prices may change.

Funding duration and practical budget

Faculty full-time Overseas research tuition is £31,350 in 2027/28; continuing fees can increase by up to 8% annually. It should not be added as a personal bill during periods fully covered by a studentship, but coverage after funded months and any project-specific bench fee need confirmation. Four-year registration alongside shorter maintenance support can leave a completion gap.

Bath's historical 2025/26 postgraduate budget totals £405 weekly/£21,060 annually, with average £230 weekly rent and a 51-week accommodation assumption. The total also includes a £37 weekly utilities line marked for off-campus accommodation. Removing that line in a genuinely bills-inclusive halls scenario gives £368 weekly, an analytical scenario rather than a new official price. Off-campus rent is illustrated at £750 monthly before bills. These figures are not 2027 contracts. The old stipend exceeds that published annual total by only £745 before relocation or immigration; this arithmetic is not a personal-surplus forecast. Deposits, full-year housing, travel between sites and upfront visa/IHS require separate provision.

Applicant Fit and Preparation

The advertisement's demonstrable-experience wording spans PDE mathematics, electrochemistry, parameter identification, embedded implementation and HPC. Applicants should present evidence across these areas and ask how breadth versus depth is assessed; the wording cannot be rewritten as “no experience required”. A master's is advantageous rather than a universal additional requirement. Strong data science alone does not demonstrate electrochemical inference or resource-constrained implementation.

A useful research portfolio would distinguish solver verification, independent experimental validation, parameter sensitivity and actual hardware cost. Evidence of a numerical method with an explicit error/conservation check, or a constrained implementation with measured limitations, can make the experience concrete. These are preparation suggestions, not extra mandatory deliverables or claims that a particular public solver is the required platform. Explain what cannot be inferred from terminal voltage and how uncertainty would affect confidence in an ageing estimate.

Transferable directions include battery modelling, BMS estimation, embedded scientific software, simulation/validation and electrified-powertrain research. The Faraday Institution's 2025 skills roundtable identifies both a shortage of trained people and gaps in sector-relevant capability; it is not a count of BMS vacancies, salary forecast or placement guarantee. IAAPS facilities and industrial links create relevant context, without promising an employer, proprietary dataset or access permission. Applicants seeking primarily materials synthesis, generic ML prediction or control tuning without physical-model development should clarify the actual research balance.

Next Actions and Questions to Verify

Before applying

Send Yuan a concise description of relevant numerical, electrochemical and implementation work. Ask which parts of the experience list are expected at entry, what model/platform baseline is intended, and whether the defined project needs an additional proposal. Resolve individual degree equivalence and English/exemption approval before the hard deadline, and check the Mechanical URSA two-choice fields. A personal award letter is not required before initial contact or entering the competition.

At interview

Clarify the target cell chemistry, reference DFN implementation, available commercial-cell data and their ownership. Ask how transport, adaptive switching and degradation are prioritised, which observables can identify the proposed parameters, and how temperature/ageing transfer will be judged. Establish the actual processor, timing/memory/precision criteria, independent validation baseline and consistency tests. Confirm co-supervision, specialist training, cell-test/technical support, HPC allocation and any industry confidentiality or publication restrictions.

Before accepting an offer

Obtain the personal 2027 stipend, funded months, tuition/indexation and bench-fee coverage, plus completion support, in writing. Coordinate the start and registration timetable with accommodation and any travel between sites. Complete the required ATAS/CAS/visa arrangements and cost the actual visa/IHS validity period. These questions determine the personal award and research conditions within the international competition route described in the October 2026 information.

Evidence and Sources

The information below is dated 18 October 2026; present vacancy availability has not been reconfirmed. The recent originals comprise two team studies and one external comparison, with the reading ranges and version discrepancies stated above. These sources support a bounded application decision, not three complete team-paper readings or independently reproduced embedded performance.

School, training and research environment

• Mechanical Engineering PhD full-time
• IAAPS purpose-built facilities
• Bath doctoral professional development
• QS Bath overall 2027 table
• Bath postgraduate research school context

Supervisor, original studies and software context

• Hao Yuan official Bath profile
• Investigation of electrochemical-mechanical-thermal coupling characteristics during lithium-ion battery charging based on an enhanced single-particle model
• Semi-supervised learning with physicochemical health indicators using an Encoder-Gaussian process framework for battery state of health estimation under dynamic discharge conditions
• A novel exact solution of extended single particle model for lithium-ion cell and its application to high capacity NMC battery
• A novel numerical implementation of electrochemical-thermal battery model for electrified powertrains with conserved spherical diffusion and high efficiency
• CPG-SPMT author-maintained research software

Funding, applications, costs and career context

• Faraday Institution roundtable on electrochemical engineering skills gap
• Current University Research Studentship Award policy and 2027 Engineering deadline
• Bath doctoral English rules
• Bath Chinese research-degree qualification route
• Bath Engineering research tuition 2027/28
• Bath postgraduate budget 2025/26
• Bath doctoral admissions FAQ 2026/27
• UKRI training grant current stipend
• UK Student visa official route
• UK immigration health surcharge
• ATAS current nationalities and subject rules
• Bath doctoral application guide and application-authorship policy