I recently came across machineering and their simulation platform iPhysics at a trade fair, and it caught my attention, so I wanted to learn a bit more about it from people who have actually used it in real projects.
I’d be really interested to hear how your overall experience with iPhysics has been in everyday use, what you mainly use it for, and how easy it is to work with in practice. I’m also curious about how quickly teams can get familiar with it, whether it fits well into existing workflows, and if there are any drawbacks or challenges that became noticeable over time.
If anyone has also worked with similar platforms, I’d love to hear how iPhysics compares and what you think its strengths and weaknesses are.
Any honest feedback or experiences would be greatly appreciated.

10 hours ago, Zerbo said:

I recently came across machineering and their simulation platform iPhysics at a trade fair, and it caught my attention, so I wanted to learn a bit more about it from people who have actually used it in real projects.
I’d be really interested to hear how your overall experience with iPhysics has been in everyday use, what you mainly use it for, and how easy it is to work with in practice. I’m also curious about how quickly teams can get familiar with it, whether it fits well into existing workflows, and if there are any drawbacks or challenges that became noticeable over time.
If anyone has also worked with similar platforms, I’d love to hear how iPhysics compares and what you think its strengths and weaknesses are.
Any honest feedback or experiences would be greatly appreciated.

From a quick look at the company website it seems very specialised to small, perhaps intricate, machine tools.

Simulation software has been around a long time and, depending upon the enormously wide range of processes (probably almost all process) that can be simulated as well as the type of simulation required, may or may not be suitable for any given application.

I don’t use iPhysics myself on a daily basis, but my employer is familiar with it, particularly with regard to virtual commissioning. The exciting part, I've learned, isn't the attractive 3D view, but whether the mechanics, control systems, material flow and operational behaviour all fit together. If you only use it to make a machine move smoothly, the benefits are limited. However, it becomes more interesting in practice if you can observe things beforehand, before it gets expensive or if you get the timing wrong. What I would advise against, though, is modelling the entire line right from the start. It makes much more sense to use a smaller model or a proof of concept in the initial phase. This will also help you to see how well it fits into the team and existing processes. What kind of systems do you want to simulate?

On 5/29/2026 at 10:28 PM, studiot said:

From a quick look at the company website it seems very specialised to small, perhaps intricate, machine tools.

Simulation software has been around a long time and, depending upon the enormously wide range of processes (probably almost all process) that can be simulated as well as the type of simulation required, may or may not be suitable for any given application.

So far, I’ve only seen the company at a trade fair and then watched a few videos of it on YouTube. The iPhysics software is supposed to be able to create a digital twin of the machine. This allows you to view the entire process on a PC. If any errors occur, you can quickly make changes before the machine is actually installed in the factory. LIke a virtual commissioning.
That would obviously be great, as it would mean less downtime or allow staff to be trained in advance. That would be interesting. The question is also how quickly you can get to master the software.

4 minutes ago, aliceinwonderland said:

I don’t use iPhysics myself on a daily basis, but my employer is familiar with it, particularly with regard to virtual commissioning.

This will also help you to see how well it fits into the team and existing processes. What kind of systems do you want to simulate?

Can one employee take care of that, or is it a whole team?
How long did it take him to get to master the software?
Did he teach himself, or are there any training courses available?

We are mainly interested in simulating automated production and material-handling systems before they are built or modified. Our goal is to create a digital representation of the planned equipment, including robots, conveyors, sensors, and control logic, so that we can test different operating scenarios in a virtual environment. This would allow us to identify potential bottlenecks, collisions, or sequencing issues at an early stage, reducing the risk of costly changes during commissioning.
Another important use case is the validation of automation concepts and PLC programs before deployment. By connecting the control software to a virtual model, we can verify that machines behave as expected and optimize processes without interrupting production. In addition, we see value in using the simulation for operator training and for evaluating future production changes, such as new product variants or increased throughput requirements. Overall, the main reason for considering iPhysics is to reduce engineering risks, shorten commissioning times, and improve confidence in design decisions before investing in physical equipment.

24 minutes ago, Zerbo said:

So far, I’ve only seen the company at a trade fair and then watched a few videos of it on YouTube. The iPhysics software is supposed to be able to create a digital twin of the machine. This allows you to view the entire process on a PC. If any errors occur, you can quickly make changes before the machine is actually installed in the factory. LIke a virtual commissioning.
That would obviously be great, as it would mean less downtime or allow staff to be trained in advance. That would be interesting. The question is also how quickly you can get to master the software.

Can one employee take care of that, or is it a whole team?
How long did it take him to get to master the software?
Did he teach himself, or are there any training courses available?

We are mainly interested in simulating automated production and material-handling systems before they are built or modified. Our goal is to create a digital representation of the planned equipment, including robots, conveyors, sensors, and control logic, so that we can test different operating scenarios in a virtual environment. This would allow us to identify potential bottlenecks, collisions, or sequencing issues at an early stage, reducing the risk of costly changes during commissioning.
Another important use case is the validation of automation concepts and PLC programs before deployment. By connecting the control software to a virtual model, we can verify that machines behave as expected and optimize processes without interrupting production. In addition, we see value in using the simulation for operator training and for evaluating future production changes, such as new product variants or increased throughput requirements. Overall, the main reason for considering iPhysics is to reduce engineering risks, shorten commissioning times, and improve confidence in design decisions before investing in physical equipment.

Staff training like a flight simulator for pilots, rather than a design tool ? A good use, but pretty expensive to set up for multiple machines.

@Zerbo Given what you’re describing, I wouldn’t simply assume that an employee could just click their way through it quickly. In my opinion, you really need at least someone who fully understands your system when all these factors come together. And you’ll also need someone who can set up the model properly. The depth of testing you intend to carry out plays a key role in determining whether one person can manage it, and we approached it back then with a limited proof of concept. Regardless of your approach, I would definitely plan for training. Otherwise, you’ll end up with nobody in your team really knowing where to start. Which control system do you mainly use?

On 6/8/2026 at 11:45 AM, studiot said:

Staff training like a flight simulator for pilots, rather than a design tool ? A good use, but pretty expensive to set up for multiple machines.

To stick with your example, would It be cheaper to provide every pilot with a real aeroplane for training?
After all, that’s exactly why they do it here, for cost reasons, and because it’s simply much easier. And if the worst comes to the worst, it doesn’t cause tens of millions in damage or cost lives.
So why shouldn’t this approach also have advantages when dealing with multiple machines?

On 6/12/2026 at 10:13 PM, aliceinwonderland said:

@Zerbo Given what you’re describing, I wouldn’t simply assume that an employee could just click their way through it quickly. In my opinion, you really need at least someone who fully understands your system when all these factors come together. And you’ll also need someone who can set up the model properly. The depth of testing you intend to carry out plays a key role in determining whether one person can manage it, and we approached it back then with a limited proof of concept. Regardless of your approach, I would definitely plan for training. Otherwise, you’ll end up with nobody in your team really knowing where to start. Which control system do you mainly use?

Even if the software itself is fairly intuitive, understanding how to build useful models and meaningful test scenarios takes time. But we don’t need to go on as if we were putting a complete beginner in that role. A small pilot project also seems like a sensible way to let the team gain experience before using it on larger machines. They know what they have to do now, so they will know what they have to do with a new Programme.
At the moment, we mainly work with one of the common PLC platforms, but we're also looking at how well virtual commissioning fits into our existing workflow before committing to it on a larger scale.

The aeroplane comparison isn’t that far off. I wouldn’t base the savings solely on the number of machines. By the time the control system and signals are functioning correctly in the model, a lot of work will probably already have gone into it. Smooth animation alone wouldn't tell me much at that stage.

I also wonder who would develop it on your end. Someone who already knows the plant and the PLC project would probably get up to speed faster. Someone from outside the project would likely only notice certain quirks once the model was already being set up. If possible, I’d therefore try out iPhysics directly on a smaller real subproject. This would quickly demonstrate how much the CAD data and PLC project actually help. It's also possible that more work will pop up in some inconspicuous spot than was expected beforehand.

Training and proof-of-concept workshops certainly help with getting started. I’d be more interested to know what happened a few months later. Is the model used again for the next project, or is there no time for it in everyday operations?

Which PLC platform do you use specifically?

I think the long-term value is the more important question than the initial proof of concept. Even if the first model takes longer than expected,that doesn't automatically mean the approach isn't worthwhile. Once a team has built up a library of reusable components and established workflows, the effort for later projects can be quite different..
I'm also not convinced that the person building the model always has to be deeply involved in the original PLC project. In many companies, simulation or commissioning engineers become specialists in the modeling side while working closely with the PLC programmers. That separation can actually make the process more scalable.
For me, the deciding factor would be how well the virtual model predicts issues that would otherwise only be found during commissioning. If it consistently catches sequencing errors, sensor placement problems, or timing conflicts before the hardware is available, then it has already justified a significant part of the investment. The exact PLC platform amtters less than having a structured codebase and reasonably complete CAD data.

29 minutes ago, Zerbo said:

I think the long-term value is the more important question than the initial proof of concept. Even if the first model takes longer than expected,that doesn't automatically mean the approach isn't worthwhile. Once a team has built up a library of reusable components and established workflows, the effort for later projects can be quite different..
I'm also not convinced that the person building the model always has to be deeply involved in the original PLC project. In many companies, simulation or commissioning engineers become specialists in the modeling side while working closely with the PLC programmers. That separation can actually make the process more scalable.
For me, the deciding factor would be how well the virtual model predicts issues that would otherwise only be found during commissioning. If it consistently catches sequencing errors, sensor placement problems, or timing conflicts before the hardware is available, then it has already justified a significant part of the investment. The exact PLC platform amtters less than having a structured codebase and reasonably complete CAD data.

You seem to have moved on from staff training ?

Your second paragraph above sounds much more useful and interesting, please elaborate.