Future project ideas

A backlog for future experiments, side projects and tooling.

A public shortlist of experiments, side projects and tools that could become working prototypes.

Project watchlist

Simulation

Aircraft de-icing simulation in ANSYS

A thermal-fluid project that combines airflow, icing conditions and surface heat transfer with a practical engineering use case.

Data

Power engineering dashboard with Pandas and Pandapower

A reusable analysis tool for grid studies, load cases and visualising power-system behaviour in a cleaner way.

Automation

Home automation orchestration layer

A small system that ties together Google Home, NFC tags and SmartThings style triggers in one place.

Embedded

Sensor logger in embedded C

A low-level build that would help bridge software, electronics and real-time data collection.

Analysis

R notebooks for statistical exploration

Useful for comparison plots, cleaner statistical workflows and broadening the tools available for data work.

Reliability

FMEA / RCA toolkit for energy systems

A structured way to catalogue failure modes, symptoms, causes and follow-up actions across thermal and process assets.

Open-source

Browser-side physics layer (extracted)

The scripts/physics/ module on this site already pulls together isentropic relations, 1-D thermal resistance circuits, Bartz heat-transfer scaling and ISO 50006 EnPI in one audited place. Pulling that out into a standalone npm package (or a Python twin) would let other portfolios + course assignments reuse the same audited formulas.

CHT validation

Bartz vs. RANS calibration on a test bench

A natural follow-on to the Siemens thesis: design a small heated-channel rig (Re 10⁴–10⁵, q'' up to ~50 kW/m²) instrumented with embedded thermocouples and a precision power meter. Compare measured h_g to Bartz-scaled predictions and to RANS k-ω SST results. Tight calibration loop for hyper-sonic / regen-cooling research.

Industrial decarbonisation

EnPI baseline-drift detector

An online add-on to the EnPI toolkit that flags when a baseline regression has lost predictive power. Uses rolling KS-test or CUSUM on residuals to suggest "this baseline needs re-fitting before reporting Q2 results". Closes the loop between metering, regression and audit-ready reporting.

Method study

Insulation cost-benefit overlay on the Pulsatorn rig

Take the thesis insulation case (40 mm ceramic blanket reverses the uninsulated heat-loss disadvantage) and extend it to a parameter sweep: blanket thickness 20–100 mm, k 0.05–0.15 W/mK, h_ext 2–10 W/m²K. Output a Pareto front of cost vs. delivered-temperature loss for the next rig-build decision.

How ideas become projects

Each idea here is a candidate, not a commitment. The filter I apply: does this idea reuse a method I already have audited, on a different evidence base? When the answer is yes — like the EnPI drift detector reusing the toolkit, or the insulation parameter sweep reusing the thesis physics — it tends to make it onto the actual project list within a few months. When the answer is no, it stays here as a thinking artefact.