D-Mould is a robotics-based solution that contributes to the continuous production of moulding machines with reduced downtime and a more reliable process resulting in decreased human risk factors and increased production output.
Experimental setup consists of Pre-Existing hardware:
- Injection moulding machine (IMM),
- Linear robot,
- Conveyor belt
and new hardware:
- Cobot 1,
- Cobot 2,
- Autonomous mobile robot (AMR),
- VLM. 878
New hardware
Both Cobots are equipped with computer vision system for parts detection, AMR is equipped with laser scanners and 3D cameras.
We‘re exporting NGSI entities to Orion Context Broker running OPC/UA IoT Agent. Instance of SQL (PostgreSQL) server is used to store historical data. Context Broker and NGSI system entities are running on local network. Database with historical data is accessable on public address dmouldramp.ddns.net through port 5432. The database is accessible on demand.
The image below shows how ROSE-AP is integrated in the Factory and connected to RAMP.
Architecture overview
Follow our video step-by-step tutorial or use installation documentation below.
Supported OS for this project is Windows 10 and Linux, however our main test environment was Windows 10 so we will include only instructions for this specific OS and we recommend using it for your testing purposes.
To run a project clone/download a git repository and follow instructions below.
Open folder x64 and start D-MouldOpc.exe. To run application you should have .NET 5.0 installed on computer.
Let the console application run in the background. The output will be similar as one on the picture below:
OPC console output
To install Docker follow the official instructions and enable WSL 2 during the installation, more WSL information.
docker-compose.yml file shows all the required information regarding ports and network configuration, as well as the environment variables for the containers that require them.
-
After installing, start the Docker Desktop program.
-
Run administrator version of PowerShell in folder d-mould and run these commands:
2.1.
wsl -d docker-desktop2.2.
sysctl -w vm.max_map_count=262144 -
Run a second administrator version of PowerShell in the same folder d-mould and run these commands:
3.1
docker compose upYou should see a similar output as on image below.
PowerShell Docker output
Let Docker run in the shell and move to the next step.
To test if everything works as it should open web browser and open the following URL: 172.28.16.1:4200, CrateDB dasboard should open.
CrateDB dashboard
Tables in dashboard are showing data from machines, further processing/querrying can be done following instructions on CrateDB website, e.g. Selecting data, etc,.
Data
Data is also visualised in our RAMP dashboard
RAMP dasboard
KPI example
Click to sample csv data example
The OPC-UA nodes are translated to FIWARE Orion Context Broker entities via mapping configuration file. The mapping is done in the following way as shown in the json file on the link below.
This is also presenting the entities available in the context broker.
• Cycle time – numeric variable, presenting the complete cycle time of a workstation
• Good parts counter – numeric variable, presenting the counter of good partis since last counter reset
• Total parts counter – numeric variable, presenting the total count of parts since last reset of the counters. It is actually presenting the sum of good and bad parts
• Target quantity – numeric variable, presenting the goal count of good parts to complete the work order
• Packaging unit quantity – numeric variable, presenting the amount of parts, which fits in transportation box. This variable is used for automatic call of the OTTO SDV.
• Sensor 1 – Boolean variable, presenting the status of the sensor at the drop zone of the conveyor. This data can be used to determine bottlenecks and unusual behavior
• Sensor 2 – Boolean variable, presenting the status of the sensor at the pick zone of the conveyor. This data can be used to determine bottlenecks and unusual behavior
• Operation mode – numeric variable, mode of the robot. It can be used to detect errors and interference of the operators
OPC-UA discovery example
The project is licensed under the Apache-2 license.