Robotics Dashboard
Solar Panel Factory Monitoring
Track and Troubleshoot Complex Processes
Robotics Dashboard
Malaysian solar panel factory
A Malaysian solar panel factory needed a way to track the hundreds of robots and processes on the factory floor. A single delay can halt the entire process and is extremely costly. Previously, the factory administrator had to read alerts in a complex table. It's easy to miss an alert when it's in a text-based spreadsheet.
THE SOLUTION
Live layout-based monitoring
I designed a heads up dashboard to monitor the robots and processes in real time. By mapping each robots' error status data to the node on a map of the factory floor, trouble spots can be easily located and resolved more quickly.
For each offline robot, give alerts indicating:
-
why the robot is offline
-
where the robot is located on the floor
-
how long the robot has been offline
CHALLENGES
Complexities and Interdependencies
Understanding all the processes and the types of things that can go wrong was unimaginable without seeing the facilities in person. The engineers understood their individual parts, but no one had a holistic understanding of the process and how the logistics of the robot processes.
CATEGORY
Factory monitoring software.
DATE
Jan 2017 - Feb 2017 3 weeks
MY ROLE
Interaction Designer
PARTICIPATORY DESIGN
Traveled abroad to be on site
Toured the facilities and saw every process
Interviewed the Chief Engineers who are on the solar factory floor, people on the line, and management
RESEARCH AND REPORTING
User Research
Interviews
Card Sorting
Understand the Data
ITERATIVE DESIGN
Multiple Iterations
Rapid prototyping
Testing
Peer reviews
Stakeholder reviews
INFORMATION ARCHITECTURE
Created an interactive dashboard
Production Wireframes
Animated prototypes
Read the floor plan
Redrew the AutoCAD drawing into SVGs
Mapped the alerts to the floor plan
Solar factory in malaysia
Wireframe iterations
THE PROBLEM
Transporter Rail
The backbone of the factory is the transporter rail, which picks up and drops off solar wafers around the factory to their different processes.
When parts malfunction, because they are on a rail, it can cause the entire side of the factory to halt. Delays are very expensive.
The factory managers need a way to track the transporters and to be given clear notification when and where a malfunction happens.
Transporters must always move or delay are inevitable
ROBOTS PROCESS THE WAFERS
Humans are only needed to keep the process moving. They need to quickly respond to the machines if there is a malfunction to make sure the assembly line doesn't fall behind a projected output.
One error delays the whole process line.
DISCOVERY
THE ENGINEERS
I discovered that the engineers understand their individual parts, but no one has a holistic understanding of the process and how the logistics of the robot processes.
I spoke with everyone who had some stake in monitoring the factory.
My Team of Engineers
THE FACTORY HAS 3 FLOORS
The chief engineers want a centralized dashboard to monitor all 3 floors of the large factory.
Each floor has a different layout because they do different processes on each floor and therefore have different robots of varied sizes.
SYNTHESIS AND INSIGHTS
OFF LINE
After talking with everyone, I understood the levels of urgency and logistical issues. Sometimes machines go offline for scheduled maintenance or other machines further down the line fail.
There are a variety of reasons a machine can go offline, not just malfunctions.
Allowing the user to easily be able to differentiate the reason for the offline is critical.
Some indicator of length of time a machine has been offline is also import.
The exact location is also required since many robots are very small and there can be many processes in one process line.
The Solution is Modular SVGs
INTIAL THOUGHTS
I thought calculating pieces/hour would be helpful. I didn't know how many machines there were in a process. I didn't know the layout of the floor or scale.
I made a good guess like staggering the Pills shapes because the Process alternates row by row and the size of the process changes. The pill allows you to easily stretch in the Javascript code.
Using the alphabet was a solid choice that lasted.
ALERT CODING
I created color codes for the different types of alert notifications for fast assessment.
TRANSPORTER ALERTS
Working out the Special GPS location alert for the Transporter.
DASHBOARD 1ST CONCEPT
Before I had any information. I just imagined from this rough blueprint what the users would want to know or do from a high level.
ITERATION 2 : BLUEPRINT MAP
Working out the Special GPS location alert for the Transporter.
GPS LOCATION
We located the API's for the GPS location that each robot has. The rail is also mapped with GPS coordinates.
I combined the GPS coordinates with the
AutoCAD blueprint. This is so I can accurately show the location of any alert. This is also easily updated in the code if the move machines around.
GPS LOCATION
I traced the AutoCAD File to make SVG's of the rail. This how we can link the robot's data
GPS and the map to the alerts location.
ITERATION 3. GPS GRID
I combined the GPS Coordinates with the
Auto cad Factory Blueprints and the letter based location labeling. This is so I can accurately show the location of any alert. This also easily updates in the code if they move machines around. Simply enter the GPS coordinates of the new location.
THE STRATEGY
ITERATION 4. A lighter touch
Once I had the function taken care of I worked on the color. Since Green indicated GOOD. There would a lot of Green. So I chose light shades of green which gives this otherwise dry dashboard a fresh feel.
ITERATION 4. The pill
I chose the pill because it's easy to read and to make. In Javascript, make an SVG rectangle, then round all the corners 90º. To change the length of the pill increase the W or H of the rectangle.
ITERATION 4. Repeat
I then applied the same design to the other floors, slightly changing the color.