We are excited to be a part of this challenge, thank you element14 community to help us be a part of this. So as to introduce us and our plans, we wanted to share our proposal; as in what we plan to do.

We are a team of students, from National Institute of Technology Warangal India, with the interest in drone technology; we aim to learn and contribute in this field.

Below is our proposal:

 

Problem

 

Industrial inspection is a key aspect of any factory or production plant. In its current state, it takes up a lot of time, man-power and non-reusable resources. It is also not something that can be avoided at any cost. The safety of the entire plant and the community surrounding it depends on the quality of the inspection that is undertaken at the plant. Let us consider the example of the oil and gas industry. A tiny leakage at any corner of a production plant could cause a large scale disaster if accidentally ignited. The same is true in places like thermal power plants. To keep such mishaps in check these industries invest heavily in the inspection sector. A large number of people are involved in manually inspecting every inch of the plant. Often, scaffolding is constructed around boilers, cooling towers, etc. so that we can reach these normally inaccessible points. This job is performed at high risk to the lives of these people. It is also very time intensive as the there is a vast area to be covered, at regular time intervals, to make sure that the plant is safe at all times.

 

Proposal

 

Our proposed system is an unmanned aerial vehicle industrial inspection system. In this, a single or a swarm of unmanned aerial vehicles, more specifically quadcopters, will be equipped with a wide range of sensors using which they will perform the complete act of inspection at much higher speed and efficiency as compared to the current methods. It would also be made completely autonomous i.e. zero human intervention at the time of inspection. Using flight planning or more generally- computational motion planning, we will be feeding in a pre-planned path to the quadcopter and it would be followed without the need of any manual control.

 

System Requirements

 

The process of inspection involves actions like detecting cracks, leakages, structural deformations and other anomalies. All of these would be achieved through the use of infrared sensors, ultrasonic sensors, radiometric cameras, normal cameras, temperature sensors etc. Our service could be of two kinds, one which involves just the collection of data and another that includes its interpretation.

The following components will be a part of our solution-

  1. A 3D printed quadcopter frame- We have already made and tested multiple versions of quadcopter frames using 3D printers. We are currently in the process of optimizing weight and material usage to the maximum extent.
  2. A flight controller- We intend to program our own flight controller using something like the STEVAL-STLKT01V1STEVAL-STLKT01V1 development kit. This would fulfil all requirements of the flight controller including an accelerometer and a gyroscope. We will be adding an inertial measurement unit to improve quadcopter stability.
  3. 4 BLDC motors and propellers to provide the thrust for the quadcopter.
  4. Sensors:
  • Infrared sensors- These will be used to detect leakages by sensing temperature difference between a pipe and its contents wherever there’s a crack.
  • Ultrasonic sensors- These would bounce off and receive ultrasonic waves from the walls of any structure and based on the intensity of the reflected wave, compute the structural reliability of a structure by relating it to the elasticity modulus of the structure material.
  • Radiometric cameras- These cameras would perform a function similar to the infrared sensors but with a higher degree of sensitivity and resolution.
  • Cameras- We intend to use computer vision to detect cracks in pipes, boilers and other structures. We will be training a machine learning model with images of pipes and this will be used for detection after images are collected by the cameras mounted on the quadcopter.
  • The RTD offered in the kit could also be used to detect minor fires that might potentially lead to accidents.
  • We are looking into and considering multiple other sensors also to collect a wide range of other parameters.

One of the major improvements in current drone inspection systems would be the complete autonomy that we offer. The UAV would take off and fly along a pre-fed path all by itself, collect the data and pass it back to a computer in real time using Bluetooth/ internet and then return back and land periodically so that its Li-ion batteries can be re-charged.

We also intend to perform certain experiments so that we can improve the power efficiency of the quadcopter. We intend to use certain basic principles of aerodynamics like the ground effect to reduce the power usage and hence increase the flight time of our quadcopter swarm. Some of these experiments would hugely involve the use of a load cell for testing.

 

Challenges

 

Some of the major tasks that we face would be making our system compliant with the rules set by the petroleum and explosives safety organizations. This is something we are highly focused on.

Another issue is the economic factor. This system would require some expensive sensors for high accuracy in results. But since it is a replacement technology, that kind of capital would be available. But we intend to make it cost as little as possible. Intensive research is being done to find the best sensors at the lowest costs possible. We are reiterating the design for the quadcopter frame as many times as we can so that the amount of material used is brought down to the bare minimum, thereby decreasing the cost of material and also increasing efficiency as the weight is decreased. We are looking into the possibilities of integrating our control circuits into the frame itself by using PCB design techniques.