English 中文
Contact Us
AIPE Explorations

Exploring what engineering has yet to prove.

Open questions, emerging concepts and AI-assisted investigations at the frontier of power and energy engineering.

Open questions Research principles
Open questions
2
In progress
0
Concluded
0
Open research questions

Questions worth investigating.

Each question states a hypothesis, what is already established and how it could be tested or falsified. At Concept stage there are no results yet.

Concept Devices Power semiconductor ageing from operational data Can power semiconductor degradation be estimated from the data a converter already measures for control, without dedicated diagnostic measurements? No evidence recorded yet Opened 8 Oct 2026 Concept Microgrids · Control · Converters Autonomous DC network stability Can interconnected DC converters keep a network stable using only local control and locally checkable design rules, without centralised coordination? No evidence recorded yet Opened 8 Oct 2026
Investigations in progress

Work under way.

Explorations with recorded analysis, models, simulations or reviews. Each card shows how much evidence exists.

No investigation is in progress yet. Pick an open question above to start one.

Research findings

Validated or refuted

Negative results are published alongside positive ones.

No investigation has reached a conclusion yet. Validated and refuted results will both be published here.

Evidence status

Every label is backed by recorded evidence.

A status names the strongest evidence recorded so far. Statuses are not a fixed sequence, and the site's checks reject any status that the evidence record does not support.

Concept
A question and a hypothesis are stated. No analysis has been completed.
Under investigation
Derivation, literature comparison or AI-assisted analysis has produced recorded output. Nothing is concluded.
Modelled
A mathematical or computational model is published with its assumptions.
Simulated
A model has been simulated and the files are published. Results describe the model, not the physical system.
Human reviewed
A named engineer or scientist has reviewed the reasoning and their conclusions are recorded. This is not journal peer review.
Experimentally validated
Measurements on physical hardware support the hypothesis within stated conditions.
Refuted
Recorded evidence contradicts the hypothesis. The work stays published so others need not repeat it.
Research principles

What each kind of evidence can and cannot tell you.

AI can accelerate an investigation, but evidence, reproducibility and human scientific judgement decide what it shows.

Hypothesis

IsA precise, testable proposal stating what would make it false.

Is notA result. Plausibility is not evidence.

AI-assisted analysis

IsDerivations, literature searches and reasoning produced with AI tools and checked step by step.

Is notScientific validation. Fluent reasoning can still be wrong, and every reference must be verified.

Simulation

IsThe behaviour of a model under stated assumptions, with files others can rerun.

Is notProof of physical feasibility. A simulation cannot be more correct than its model.

Human review

IsA named engineer or scientist checking assumptions and reasoning, with the outcome recorded.

Is notPeer review or experimental validation. A reviewer can only check what was shown.

Experimental evidence

IsMeasurements on physical hardware, with conditions, uncertainty and data published.

Is notUniversal truth. It holds for the conditions tested and can be overturned by better evidence.

From question to artifact

Explorations explain the research. The Hub holds what it produces.

When an investigation produces a dataset, model, design, tool or workflow that others can reuse, it is published to the AIPE Hub with its own maturity. Publication never marks an artifact as validated; the exploration keeps the reasoning, limitations and evidence behind it. Explore the Hub →

  1. Exploration
  2. Hypothesis
  3. Modelling & simulation
  4. Human review
  5. Reusable artifacts
  6. AIPE Hub
Contribute to an exploration

Review it, challenge it, extend it.

Every exploration is a Markdown file in the website's public repository, written to a shared template. Suggest changes on GitHub, or contact us to collaborate on a larger investigation.

Open a GitHub issue ↗ Read the template ↗ Discuss a collaboration
01

Review or challenge
Check an assumption, a derivation or a reference. A well-argued objection is a contribution, and negative or inconclusive results are published.

02

Extend the work
Contribute a model, simulation, dataset or measurement. Each piece of evidence links to its files and states its limits.

03

Propose a question
Suggest a question worth investigating, with why it matters and what result would answer it.