dtaas-wt-swap

This repository contains a water-tank model swap experiment for DTaaS. The experiment uses the Maestro co-orchestration engine for FMI-based co-simulation with the Model Swap feature including a FaultInject extension. The repository is a Docker-less version of the experiment at https://github.com/lausdahl/fmiSwap for use on the DTaaS platform.

The model swap mechanism demonstrated by the experiment is detailed in the paper "Dynamic Runtime Integration of New Models in Digital Twins" [H. Ejersbo, K. Lausdahl, M. Frasheri, L. Esterle] presented at the SEAMS 2023 conference. The water-tank experiment is further detailed here: https://arxiv.org/abs/2304.07328.

Install in DTaaS workspace:

git clone git@github.com:hejersbo/dtaas-wt-swap.git <WORKSPACE>

or if folder structure already exists and is non-empty:

cd <WORKSPACE>
git init
git remote add origin <this-repo-url>
git pull
git checkout main -f
git branch --set-upstream-to origin/master

Run in DTaaS workspace:

cd <WORKSPACE>/digital_twins/water_tank_swap
./wt-run.sh
./post-process.sh

The post-process.sh runs an experiment specific python plot of outputs. Plot results are in folder post.

Prerequisites:

If not already installed on the DTaaS platform, maestro requires the Java JDK 11 to be installed:

sudo apt-get install openjdk-11-jdk

In addition, these python packages are needed for post processing:

pip install pandas matplotlib