Talk:SimScale

Untitled

Hello! I wrote this page about SimScale, which is a web-based simulation platform with a free plan. It integrates open source solvers like OpenFOAM, Code_Aster or CalculiX. I got several comments on it and I would really appreciate your help to fix them. I'm open to edit the page accordingly.

1. This article is an orphan, as no other articles link to it.

Although SimScale was mentioned in some pages, they didn't have the link to it (the brackets). I added them now, as well as new ones for relevant pages and this error is not applicable anymore. How could I remove it?

2. This article contains content that is written like an advertisement.

I removed a sentence that could have been interpreted as advertisement. Do you still think it has this problem? If this is the case, could you please point out which paragraphs?

3. A major contributor to this article appears to have a close connection with its subject.

I disclosed the fact that I am an employee at SimScale like it is written in the rules. But other pages about companies don't have it. What should I do in this case?

The purpose of this page is to give impartial information about the company. And given the fact that it uses open source software, it is available for free (for public use) and has a growing CAE community, I think it is relevant to be on Wikipedia.

I would very much appreciate your feedback! — Preceding unsigned comment added by Andreea Zaha (talk • contribs) 13:12, 14 March 2016 (UTC)

Edit request: expand Features section

I work for SimScale GmbH (disclosure on my user page per WP:PAID), so I am requesting this change rather than editing directly.

The current Features section describes only the thermal module. I'd like to propose replacing it with the expanded version below, which covers the platform's simulation solvers and AI capabilities. All new claims are cited to independent trade publications (Engineering.com, Digital Engineering, AEC Magazine, World Pumps, Design News). The existing thermal sentence is retained unchanged. All references are defined in full within the proposed text, so a few duplicate sources already cited elsewhere in the article. Should these be merged? Where the text mentions the company's own marketing names ("Physics AI", "Engineering AI"), this is attributed as such rather than stated in Wikipedia's voice.

Proposed text:

Proposed draft

Features

SimScale runs in a web browser, with simulations computed on cloud servers rather than the user's local hardware.

Simulation

The platform's solvers cover structural, fluid, thermal, and low-frequency electromagnetic analysis. Finite element structural analysis is provided through open-source codes, as well as the Marc solver, which was added in 2025 through a partnership with Hexagon AB as an option for advanced nonlinear analysis. Marc has been owned by Cadence Design Systems since its 2026 acquisition of Hexagon's design and engineering business. For fluid dynamics, the platform combines finite-volume solvers based on open-source codes with Pacefish, a proprietary GPU-based lattice Boltzmann solver developed by Numeric Systems GmbH, used for external aerodynamics applications such as wind load and pedestrian wind comfort analysis. The thermal module allows uncoupled thermo-mechanical, conjugate heat transfer and convective heat transfer simulations. Low-frequency electromagnetic simulation, including magnetostatics, is provided through a solver from partner EMWorks. In 2025, the company partnered with AI Engineering GmbH to add a smoothed-particle hydrodynamics (SPH) solver to the platform.

Artificial intelligence

In 2023, SimScale introduced AI-based surrogate models that are trained on simulation results and predict physical behavior without running a full simulation, using graph neural networks. The company positions these models, which it markets as "Physics AI", as a complement to conventional solvers for rapid design-space exploration rather than a replacement.

In March 2025, SimScale released a pre-trained foundation model for centrifugal pump simulation, developed with Nvidia's PhysicsNeMo framework and announced at Nvidia's GTC conference. The model was trained on simulations of more than 50 pump designs drawn from the platform's public projects repository.

The platform also includes AI assistants, marketed as "Engineering AI", intended to automate simulation setup and workflow tasks.

Happy to adjust wording or trim anything reviewers consider promotional. Thank you. GoWithTheFlow (talk) 13:22, 17 July 2026 (UTC)

Hi GoWithTheFlow, and thank you for your suggestions. I am declining this edit request because Digital Engineering 24/7 is not a reliable source. The articles from Digital Engineering say that they are based upon SimScale's press releases; indeed, the article about the lattice Boltzmann solver closely resembles the press release on the company's website. WP:NEWSORG cautions against relying on "news" articles which are are really a reprint of a press release, as such articles are considered churnalism. I also have concerns about the reliability of engineering.com. For those reasons, I must decline this edit request. We appreciate your understanding. Altamel (talk) 16:50, 7 September 2026 (UTC)

Hi @Altamel, thank you for reviewing, I appreciate your time on this. I understand your concerns about DE24/7 - would you be willing to review an amended edit which does not rely on this source? I would be happy to look for alternative sources, or drop the relevant sentences if none can be found. I still feel there is a lot of scope to improve this section, which only currently describes a single part of the software (rather arbitrarily). Thanks, GoWithTheFlow (talk) 18:51, 7 September 2026 (UTC)

No problem. I have no objections if you wish to rewrite and submit another draft, but with the edit request queue backlogged at 800+ requests, I can't promise to review a resubmission. You may have better luck if you place another {{edit COI}} template so that other reviewers can take a look once you are ready. I would advise you to very carefully vet any sources that you are citing. DE24/7 is only one example of the low-effort content farms out there, and with the proliferation of AI-generated content, keep in mind that any articles which are generated in part or in whole by AI are presumed to be unreliable by our policies. If you can only find marginal sources, it would be better not to include the sentences they are meant to support. Best, Altamel (talk) 21:19, 7 September 2026 (UTC)

References