Over the 2022-2023 and 2023-2024 academic years, I was a part of the UCSB Apical Robotics Capstone Team creating a large scale vine robot for use in industrial pipe inspection. This project was part of a research contract with Bechtel, where Bechtel provided funding and created a test fixture in Houston. In my first year working on the project, I helped create the sensor mount at the front of the vine robot and a fan switching mechanism to redirect airflow to and from the vine robot. After the other members of the team graduated in June 2023, I assumed leadership of the team and a new cohort of Capstone students for the 2023-2024 academic year. Following a partially successful on-site test at a Bechtel Test Site in May 2023, we were able to refine out design and successfully inspect the test pipe specifications proposed by Bechtel in April 2024.
Original Tasks:
Our goal for the academic year was to build a 3 foot diameter vine robot that could navigate the 120 foot horizontal and the 40 foot vertical section. This was eventually shortened by Bechtel to a 60 foot horizontal and a 20 foot vertical section due to manufacturing constraints. The vine robot would include the fabric body, base station, pressurization methods, and tip mount for video streaming.
Over the 2023-2024 academic year, we significantly deviated from last year's design (see below for more detail). Rather than using a complex sensor mount that constantly translates at the tip of the vine robot during growth and retraction, we used a smaller vine robot to eject a borescope and stabilizing pneumatic pouches to the end of the pipe. During the inspection process, we pulled the entirety of the vine robot out of the pipe, streaming video from the borescope to the operator. This design was able to successfully inspect the entirety of the pipe.
As the leader of the Capstone team, my role on the team was to ensure we met our class and sponsor deadlines. In addition to quarterly updates with the Bechtel Team, I also reached out to an unfilled US Navy HVAC inspection project and we were able to run some preliminary on-site testing at Port Hueneme Naval base before flying out to Houston.
My technical responsibilities also involved developing and building the pneumatic sensor mount and managing the design of the base station.
I developed and built the pneumatic tip mount used to stabilize the borescope inside the pipe. This tip mount is pressured using pneumatic lines from the base of the vine robot and serves to center the borescope inside the pipe as well as prevent the borescope camera from free-fall in the vertical section of the pipe. The chambers are constructed using heat sealed mylar with a layer of sewn ballistic nylon to provide abrasion resistance during the inspection section.
The two-sectioned design was to test a pneumatic steering concept, but one section was removed due to uncontrollable air leakage during on-site testing.
See above image for scale (each chamber is 18 in wide, 36 in long)
Each of the three chambers surround a central tube used to limit the movement of the borescope camera. This allows the borescope to be completely constrained inside the pipe once the chambers are inflated.
The diagram to the right demonstrates the construction of the tip mount chambers.
The base station inflates the vine robot and passes through material using a gasket mechanism. Two micronel blower fans inflate the vine robot, with an airtight zipper providing access to the vine robot material.
I chose the fans, other electrical components, airtight zipper, and vine robot interface. I assisted in the gasket prototying and construction.
Over the 2022-2023 Academic Year, we developed a prototype primarily based on previous literature. This included a sensor mount that translates with the tip of the vine robot and a fully enclosed base station. However, this presented significant challenges on the scale of the Bechtel requirements. As a result, the sensor mount we built was not able to traverse further than the 90-degree bend in the pipe.
This was the sensor mount I designed after 3 months of iterative prototyping and testing. The sensor mount is a platform for the cameras and pressure sensors at the tip of the vine robot. Since the "tail" of the vine robot moves at twice the speed of the body, the sensor mount needs to constantly pass through vine robot material to stay at the front of the vine robot. I discovered that unlike previous tip mount designs, the low pressure of the large diameter vine robot was not sufficient to support a sensor mount without external supports. I was primarily prototyping using a 1-foot vine robot to simulate the larger vine robot. A version of this prototype can be seen below. The sensor mount was eventually refined to be lighter and less mechanically complex.
I also created a fan mechanism to switch the direction of airflow from our Pelonis blower fan from inflating to deflating the vine robot. This fan mechanism connects to the base station using quick-connect dryer vents and is contained inside of a plastic box with removable sides for easy transportation. The servo controlled mechanism allows us to switch the airflow in less than 1 second, giving us more precise control over the pressure of the vine robot than relying on the pwm control of the blower fan.
In May of 2023, we flew to Houston to test our vine robot in a test fixture at a Bechtel construction site. While we performed limited testing in Santa Barbara, we were not able to simulate the conditions of a 40-foot pipe with a 20-foot vertical section(shortened from original specifications). I assisted in the logistics of transporting our base station and various subsystems to Houston, as well as planning the trip. Our robot was able to grow through the length of the pipe unconstrained(without a tip mount), but failed in the bend once we assembled the tip mount at the front of the robot.