Open models
Execute the engineering assumption.
MATLAB and Python references for battery thermal response and DC-link sizing, with tests and limits in view.
Open model libraryElectrical R&D · Power Electronics · Agentic Engineering
Mohammad Rezwan Khan
I turn battery thermal behavior, power-conversion constraints, and QA requirements into auditable engineering decisions.
Mohammad Rezwan Khan is an Electrical R&D Engineer and an Aalborg University PhD in Energy Technology. His work connects battery thermal systems, BESS QA/QC, FAT/SAT readiness, grid-forming control, and MATLAB/Simulink model validation.
- Battery systems
- Power electronics
- BESS assurance
- Agentic AI
Inspectable work
Start with something you can examine.
Models, quality artifacts, and AI workflows carry visible assumptions, boundaries, and source paths.
BESS QA/QC
Trace acceptance evidence.
A sanitized path from requirements through FAT/SAT, NCR closure, commissioning, and handover.
Open artifact libraryAgentic Engineering
Keep the agent accountable.
A curriculum draft for tool use, automated checks, visible failures, and explicit engineer approval.
Review the syllabusOpen engineering models
Equations are more useful when they can be executed and challenged.
The catalogue connects audited open-source projects and focused model views to pinned code, tests, research context, and explicit limits.
Battery · converters · BESS control · MATLAB/Simulink
MATLAB Simulink Energy Lab
RC → SOC-EKF → electro-thermal → BESS control
Explore the validated labBattery thermal · MATLAB/Python reference
Battery thermal response
Cth dT/dt = I²R − hA(T − Tamb)
Inspect thermal assumptionsPower electronics · MATLAB/Python reference
DC-link ripple and capacitance
Cmin = 2E / (V₁² − V₂²)
Inspect sizing assumptionsBattery degradation · analysis software · Python
Battery Cycle-Life Analyzer
capacity fade → model comparison → bounded EOL
Compare degradation modelsBattery thermal · research translation · Python
Battery Thermal Modeling Notes
source → equation → interval → engineering decision
Trace the thermal evidenceGrid storage · research benchmark · Python
VoltRL
statet + informationt → policy → value
Inspect the benchmark protocolBESS control · validated MATLAB/Simulink model
Grid-tied and grid-forming control
P, Q, V, f → modes, limits, verified transitions
Review the validated controllerGrid services · executable playbook · Python
Smart Grid Storage Playbook
P, Q, f, V, SoC → constraint-aware dispatch
Explore grid-service limitsBESS delivery · QA/QC toolkit · Python/JSON
BESS QA/QC Toolkit
requirement → evidence → gate → acceptance
Inspect the readiness gateEngineering decisions
What had to be decided, and what evidence made the decision possible.
Employer-and-date context stays visible; confidential project data stays private.
Utility-scale ESS/EPC review
- Situation
- Bidder assumptions and interfaces needed a common review baseline.
- Action
- Structured sizing inputs, technical responses, and risk ownership.
- Result
- A decision-ready review trail before procurement and execution.
BESS QA/QC and acceptance
- Situation
- Acceptance evidence was distributed across suppliers and milestones.
- Action
- Linked criteria, FAT/SAT evidence, NCR status, and commissioning records.
- Result
- A traceable handover path with visible ownership and closure status.
Human-supervised AI
AI should accelerate engineering work without obscuring responsibility.
The workflow exposes the objective, approved tools, generated artifacts, checks, stop conditions, and accountable reviewer.
Bounded input
Define what the agent may and may not decide.
Record the objective, system boundary, available evidence, prohibited actions, stop conditions, and approval authority.
Illustrative workflow scenario
Illustrative scenario: BESS thermal safety review
-
Proposed — The engine suggested a conservative cooldown profile and a revised current envelope for a candidate pack design.
-
Illustrative check — In this hypothetical scenario, checks flag two edge cases outside acceptance bounds, including edge-zone hotspot growth in case 4; no test was run and no evidence package is implied.
-
Engineer review — The engineer withholds acceptance because the illustrative residuals do not meet safety bounds and requests revised boundary conditions and another validation run.
Illustrative outcome: Human approval remains required; this is not an executed pilot, test result, or field outcome.
See the illustrative workflowPublished research
Research earns another life when its assumptions can be inspected in code.
The DOI establishes the scientific record. The implementation exposes selected equations, assumptions, tests, and engineering limits.
- DOI
- Question
- Equation
- Implementation
- Test
- Known limits
Research software
Inter-cell heat exchange → deterministic 3-D graph evidence
Inspect ThermoWeave-3DThermal systems
Spatial temperature evidence → reduced-order sensitivity model
Verify publisher recordState estimation
Measured behavior → online identification and SoC estimation
Verify publisher recordBESS control
Operating modes → controller requirements and scenarios
Verify publisher recordTechnical conversation
Bring one model, requirement, or review bottleneck.
Send the system boundary, available evidence, and decision deadline. I will reply with the most relevant model, case study, or technical next step.