Hello,

I am developing a digital twin of a power plant. The twin, along with 
associated optimization algorithms, will be implemented in Matlab. My work is a 
part of a larger project, where several similar-ish tools will be integrated 
into same toolbox. The project has suggested StreamPipes to be used as a common 
platform for this shared toolbox of all sub-projects.

Now I would need to confirm whether this kind of Matlab digital twin can be 
integrated into StreamPipes or not. Unfortunately I do not have the skills to 
evaluate such a thing. After three days of trying, I have been able to install 
StreamPipes and I have run the tutorials to have a vague feel for what the 
software does.

IF I have understood correctly, I would need to create a custom pipeline 
processing element that takes input signals, calls _somehow_ the algorithms 
written in Matlab, and sends the outputs to sink. Is this correct? This part 
would be some sort of Java-implementation(?), along the lines of examples that 
are given in Developer Tutorials. As I have zero experience in Java, this 
interface would need to be either super simple to do, or done by someone else. 
Once done, it would need to be fixed for given inputs and outputs - to enable 
working with the Matlab-part while not worrying about the Java-part.

My assumption is that the best way to integrate the Matlab-part would be by 
compiling it as an executable. Running Matlab executables requires an 
installation of standalone Matlab Runtime. How to have Matlab Runtime running 
in StreamPipes in order to run the integrated Matlab executables, I have no 
idea. I actually don't even know if any of what I've written so far makes sense.

I would greatly appreciate if someone would take the time to answer "Yes, this 
kind of thing can be done!" or "No, it would require at least a few weeks of 
work." to the question of integration of Matlab to StreamPipes. If added 
information is needed from the Matlab-side to determine if integration is 
possible, I can be in contact with Matlab's support.

Markus Neuvonen | PhD student , M.Sc.
University of Oulu | Intelligent Machines and Systems
Phone: 040 584 3079
Room: TF318

Reply via email to