---
title: "Engineering with AI: our talk at Viennahydro 2026"
author: "Erlach & Erlach GmbH"
date: 2026-09-28
language: en
canonical: https://www.erlach-erlach.com/en/journal/viennahydro-2026-talk/
license: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)
---

# Engineering with AI: our talk at Viennahydro 2026

*On 20 November 2026 Philipp Erlach speaks in Laxenburg near Vienna on “Engineering with AI: Making Errors, Keeping the Lessons”: what an AI can compute in turbine engineering and where people are still needed.*

Erlach & Erlach, Journal, 2026-09-28. Web page: [/en/journal/viennahydro-2026-talk/](/en/journal/viennahydro-2026-talk/)

Viennahydro is the hydropower conference in Laxenburg near Vienna. In 2012 we presented the Pelton turbine System Erlach to the professional community there. On 20 November 2026 we return with a different question: what can artificial intelligence do for hydropower?

## Computing gets cheap, judging stays costly

Josef and Philipp Erlach computed two distributors for Pelton turbines with an AI: on a notebook, in 31 runs, just under six hours in total. Without nozzles, the results match the values an institute obtained in 2020 for a comparable distributor.

The effort lay elsewhere: in finding the errors. We counted 17, and in the talk we show who found them, the designer, the agent or a second model acting as reviewer.

## A good AI system is built like a good team

Our experience from just under eight weeks: it needs a second reader, a duty to cite sources, a check against known values and the option to say: I don't know.

The figures apply to this one case. We chose it because we know our own turbine and could check the results against earlier measurements and calculations.

*After the event in Vienna we will publish a report on it here, together with the full content of the talk.*

[Go to Viennahydro 2026](https://www.viennahydro.com/)