During my time at ATCO EnPower, I helped develop many different tools to improve workflows, such as fault-tolerant servers and dashboarding software, but in particular I helped automate many different spreadsheets. Each of these spreadsheets required manual entries that were both time-consuming and inconvenient. The water report was the project that made the largest positive impact.
The Water Report is filled out each month by operators at one of EnPower's renewable sites so Operations Technology can monitor and track various metrics to ensure consistency and reliability. The problem was that the spreadsheet could only be filled in manually, which involved travelling from a particular computer running outdated software to a newer computer. This process was taking 4-6 hours per month to fill out each piece of information in the sheet.
I was tasked with finding a way to automate the spreadsheet so it would take less time to fill. Instead of using the outdated system to get the values, there were now PI tags in place of the old system. As a side task, I had to find each of the missing PI tags so a PI specialist could create them.
Using over 30 PI tags and various Excel macros, I automated the report to fill in all the numbers from PI with the push of a button. This automation changed the time to fill in the monthly tallies from 4-6 hours to exactly 2.6 seconds for the entire month. Combined over a year, this change saved operators around two weeks of straight manual entry.
This automation not only saved time and money but made the lives of the operators easier and made the entire process foolproof for the future.
My mecanum-wheeled vehicle project was based on research I had been doing on theme park rides. I was incredibly interested in how ride vehicles were able to turn without needing to steer the front wheels, allowing for quick turns at almost any instant. That is when I came across mecanum wheels, and I immediately wanted to develop a theme-park-esque ride vehicle.
Before I began building, I came up with preliminary designs. This design featured a flat rectangular body with rounded ends, housing a very large passenger section with space for the motors within the rounded ends. There were a couple of iterations of the model, but even as the scale increased and the passenger space shrank, the preliminary model could not fit all of the materials needed without greatly expanding its size, which caused me to pivot to the revised model.
After the last complete model failed due to size constraints, rather than building and ordering the pieces to fit around the shape, I redesigned the model to specifically fit each of the pieces I required in my vehicle. I created a thin, flat rectangular plate with specific slots to fit each piece of the vehicle. I designed it this way so I can expand it in the future.
To test the model and fully develop my understanding of how mecanum wheels can be used for ride vehicles, I wrote a simple script that makes the robot go forwards, backwards, left, right, and stop at random in quick movements. For future expansion, I am planning to add a LiDAR sensor to scan the environment around the vehicle so I can code it to be autonomous. I am also planning to add two more layers on top of the current layer to develop a back-projection system. I have already developed a head and projection system; I just have to program the face animation.
A trackless dark ride vehicle has to do two jobs at once: move guests precisely through a scene, and be part of the show itself. My current vehicle proved out the mecanum drive. The next version, affectionately called SPECTER, is now in design: an RC-scale, 3D-printed model of a trackless ride vehicle with two rows of four seats and a Pepper's ghost display built into the dashboard, so the vehicle carries its own show effect. I'm test-printing a small version of the body now. Next come the full-size print, drive motors, LiDAR, and the wiring and autonomy code. I'm still deciding whether it will follow a scripted show path or roam freely.
During my co-op at Stumpcraft, I made many proposals for updating outdated workflows that were bottlenecking the production line. The Custom Puzzle Automation was the most successful workflow revision I worked on.
Before my revision, the Stumpcraft custom puzzle process involved various people individually pulling the images being used for the custom puzzles from Shopify. After getting the images, they had to rename each image with a specific name and put it in Dropbox. Then they had to individually make labels and inserts for each custom puzzle and rename each of those to specific names. Each month, these custom puzzles cost people days of work to make and rename the pictures, labels, and inserts. I thought it was a huge waste of time, so I developed an automated flow that does most of the work for the person.
To fix this workflow, I automated pulling the images and image names into a Google Sheet. In the sheet, I was able to sort the images by name and generate names for the pictures, labels, and inserts. Then using the picture names in the sheet and over 800 lines of C++ code, I renamed every image with the correct title and put each one in the correct folder. Finally, I bulk-upload the sheet into a Canva template, which makes all the labels and inserts in a matter of seconds. My revised workflow saved weeks of work throughout the year!
Design a free-roaming character that does a real task for a guest, pouring exactly one cup of water, while staying in character and moving safely around people.
Hydrate drives on a two-wheel base and uses servos to move and express itself. A metered pump measures out the pour, microphones let it turn toward whoever is speaking, and its face is a small Pepper's ghost display: a screen reflected in a hidden angled acrylic panel.
Design and simulation. The character concept is done, and I'm working through the parts list and building a simulation model before ordering anything. The physical build is next.