Pipe Crawling Robot

This project is sponsored by ARPA-E and is associated with the research work I do in the BioRobotics lab at Carnegie Mellon. The goal of this project is to create a pipe crawling robot for natural gas pipelines to perform inspections and repairs inside the pipes. The robot must be able to navigate twists and turns as well as vertical sections of pipe without getting hung up on service connections. Note, due to information confidentiality and intellectual property protection, I will not be able to go into great detail about this project however, it is my goal to summarize the scope of my involvement.

Background

Old natural gas pipelines have been underground for decades and many of them have gone without inspection for a significant amount of time. There are many issues that can occur with these pipes such as corrosion, erosion, cracks, and leakages and to fix these issues the pipes need to be dug up which can be very slow and costly. Our (Carnegie Mellon Robotics Institute) solution involves the robotic construction of a novel, structurally independent, self-healing, and self-reporting pipe material that is created within the existing pipe and has minimal to no disruption of gas services. An innovative modular robotic platform is being designed and created to perform three functions: inspection of the old pipe, construction of the new pipe, and inspection of the newly constructed pipe. The new pipe material will be a tough and durable polymeric material.

Incorporated self-healing and self-reporting functionalities coupled with the inspection components of the robotic platform can eliminate manual efforts to detect and repair damage in the new pipe material by providing real-time data and visualization. This approach, encompassing state-of-the-art robotics and smart material technology will enable a new pipe rehabilitation solution that will lower rehabilitation and maintenance costs for over 60,000 miles of legacy natural gas pipelines.

My Involvement

This past semester I was tasked with designing the mechanism that would allow the robot to climb vertically inside the pipe. My solution needed to incredibly strong and robust in order to provide enough traction to support the weight of the robot, the data tether, and the new pipe material feed hose. As the robot gets further into the pipe the tether becomes longer which increases the weight and the frictional resistance as the tether drags along the pipe floor and around pipe bends. Additionally my lab wanted my solution to be retractable so when the robot traverses level pipe there would not be the unnecessary drag of my device pushing against the pipe wall to create the required normal force for climbing.

Brainstorming

In order for a robot to do vertical climbing inside of a pipe that has relatively smooth walls there must be a pushing force on the pipe wall that suspends the robot in place. Considering that the robot already has a set of powerful drive motors I decided to create a passive wheel set that would extend above the top surface of the robot and press against the pipe to supple this required normal force. I started by brainstorming ideas for a lifting mechanism, my ideas are shown below.

I decided to pursue a scissor lift style design with one end fixed and one end free to slide on a track. A choose to use a linear actuator to supply the motion to my device as they are strong, non back-drivable, and compact. Space issues played a big role in this project as the pipe diameter is only 12 inches and the robot body already takes up a large amount of space. Additionally the length of the robot sections needed to remain short, no more than about 10 inches to insure that the robot would not get stuck in pipe elbows. Therefore, my design needed to be incredibly space efficient.

Height Testing CAD

One of the most important specifications to know when sourcing a linear actuator is the stroke length. In order for me to know what stroke length is needed, I first needed to determine how high my wheels would need to go to make contact with the pipe wall. Because this robot is operating inside a pipe, the distance from the top of the robot to the pipe surface will change as the distance from the center increases. I created a CAD model that would determine the delta height based on the distance from the robot center. In the picture to the right you can see the pipe, the robot body, and example height measurements. The robot body is accurately located in the pipe based off of the driving wheel dimensions and motor locations within the body. Obviously, the closer to the center, the more room we would have to work with but we cannot make contact with the pipe directly in the center due to service connections. If the pressure wheels fall into a service connection the robot would get stuck.

Stroke Length Testing CAD

Once we made a decision on how far from the center the pressure wheels should be I was able to get the height measurement and could move on to calculating stroke length. I created a stroke length testing CAD model that had mock linkages I could adjust the length of, a mock wheel that had the appropriate dimensions of the smallest wheel that was strong enough for this application, and a height stick for a visual reference of how high the wheel needed to be extended. In this model the wheel was represented as a square just because it made the mates easier to define. I then adjusted the linkage lengths until I achieved the height I needed and until the total length of the device remained within the length of the robot. Due to the fact that the wheel had a non negligent diameter the travel distance of my device was relatively small at only 1.25 inches. Therefore I concluded that sourcing a linear actuator with a 2 inch stroke length would be more than sufficient.

Initial Device

Here is my initial CAD design with a linear actuator that has an appropriate stroke length and is strong enough to supply the desired normal force to the pipe wall. As you can see this linear actuator, although small, is still not small enough to fit underneath my mechanism in the lowered position. Space was extremely tight on this project.

To solve this issue I did some more digging and was able to find a special type of linear actuator called a bullet linear actuator which is way more compact than traditional ones. Additionally there was some space below the top surface of the robot that I was able to use to nest part of the linear actuator between other electrical components.

Final Device

Here is my final design with the new compact linear actuator positioned below my wheel extension device and slightly inside the electronics area of the robot. The only way to fit the linear actuator between existing robot components was to attach it to my mechanism slightly off center. Although this is not ideal, the offset is not extreme, the motion is not constant, the stroke length is relatively small, and there are very sturdy linear carriages that only allow motion in one direction. Therefore I felt that this design decision given the space limitations was acceptable.