Skip to main content

Building a DIY Mini Gas Turbine: Part 1 - The Compressor

Introduction and Motivation

Hello! My name is Hasitha Senevirathne and I have just completed my 3rd year in Biomedical Engineering at the University of Glasgow.

I originally had the idea for this project about two years ago during my summer holiday. As a DIY and electronics enthusiast, one of my favourite hobbies is taking apart broken devices and household appliances to obtain useful components for my projects. I hate seeing defective and unwanted items get discarded while they could still have functional value.

Thus, I started this project both as a challenge to myself and a challenge to prove that even the most complex engineering constructs such as the gas turbine can be simulated using recycled and repurposed parts of everyday objects. It will demonstrate the basic operating principles of a commercial gas turbine but on a miniature scale.

Overview

The final prototype of the turbine will consist of 3 main segments along the main shaft: the air compressor, the combustion chamber, and the power generator.

As this is a fairly complex and time-consuming project which I am undertaking alone during my holiday periods alongside internships, summer projects, and other commitments, I will tackle each component in turn. Hence, this article will primarily focus on the starter motor characteristics and the compressor design, which I believe are the most challenging, but crucial elements.

Fig 1: Overview of components along the main shaft (solid lines) and auxiliary shaft (dashed lines)

Fig 1: Overview of components along the main shaft (solid lines) and auxiliary shaft (dashed lines).

Starter Motor Characteristics

In an industrial gas turbine, the starter motor is usually incorporated directly onto the main shaft and is not used once the turbine reaches its operating revolutions per minute (RPM). However, for the sake of practicality in this miniature model, I have decided to provide its own auxiliary shaft (connected via reduction gears to the main shaft) in order to amplify the torque. This will make it easier to drive the compressor to achieve a reasonable RPM without needing to use a very high-power industrial motor.

Additionally, the starter motor will need to be powered externally and employed continually during testing until a final RPM generating enough power to self-sustain the starter motor can be achieved. The function of the starter motor in the final prototype will be controlled remotely using Arduinos sourced from RS Components via the Student Project Fund.

Initial Compressor Design

The image below summarizes my current progress to date on the prototype including cut/uncut compressor blades, housing components, and modified bearings from a fidget spinner.

Fig 2: Summary of prototyped components, including blades from an initial compressor which were far too small (middle column) compared to the current housing and uncut blades (left and right columns)

Fig 2: Summary of prototyped components, including blades from the initial compressor that were far too small (middle column), compared to the current housing and uncut blades (left and right columns).

When deciding on the size of my miniature turbine, I decided to work based on my material limitations. Since I was using recycled aluminium sheets flattened from food and drink cans, I could create fairly large compressor blades. However, I realized that mounting them and sealing the housing became more difficult as the size of the blades increased. In fact, the biggest challenge I faced throughout the compressor design was not the blades themselves; it was housing them without interference and creating an air-tight seal at such a small scale.

Initially, I designed the blades to fit inside the hollow metal casing of an old motor that I had taken apart previously. I cut 8 compressor blades to fit this diameter (2 for intake and 6 for compression) using polar graphic paper with 21 spokes as a template.

Fig 3: Cut blades of the initial compressor

Fig 3: Cut blades of the initial compressor with 21 spokes each.

Despite my novice attempt, the completed blades were highly accurate and fit the housing perfectly. However, I soon encountered two major problems:

  1. Mounting the shaft on such a small diameter housing became incredibly difficult while maintaining a perfectly straight alignment since the tolerances required were virtually unachievable using the tools available.
  2. I had only considered the rotor blades of the compressor and had not considered the space or the means to mount any stator blades.

Final Compressor Design and Calculations

Despite the setback, I used the experience from having successfully hand-cut the full set of blades to transfer to a larger diameter housing; making careful measurements and considerations for mounting the shaft, bearings, and stator blades.

views of new uncut compressor stator blades

Fig 4 & 5: Separated (left) and partially assembled (right) views of new uncut compressor stator blades and housing.

Several improvements were made from the initial prototype including:

  • Split housing rather than a full cylinder housing to improve the ease of component mounting
  • Flaps extending from the stator blades and corresponding notches on the housing to mount them accurately with even spacing
  • Additional space on the front and back to allow space for mounting more blade stages if required

Industrial gas turbine compressors typically operate at 3000 – 4000 RPM and can contain up to 20 stages (rotor and stator blade pairs) with each stage producing an approximate pressure ratio of 1.3. Therefore, the overall pressure ratio of a compressor can be calculated approximately by raising 1.3 to the power of the number of stages. For example, in a 12-stage compressor, the final pressure ratio of the inlet to outlet would be 1:1.312 ≅ 1:20.

Since my prototype does not need to provide a net positive output, I do not need as many stages of compression as an industrial turbine. Additionally, although it is fairly simple to add more stages to my compressor to increase the pressure output, the difficulty arises from sealing the housing containing many blade stages to maintain such a high-pressure ratio. Thus, if my compressor was ideally sealed, it can theoretically achieve a maximum compression ratio of approximately 1:5 with 6 stages, which can be further increased with higher RPM. I have also left space in my compressor prototype to add up to 4 additional stages of compression if required in order to achieve my end goal in successive iterations.

Final Remarks and Next Steps

At present, a lot of work still remains to complete the prototype of the compressor including cutting the rotor blades and mounting all the internal components including bearings. Then, it will need to be tested using a differential manometer to identify the maximum pressure ratio that can be achieved and troubleshoot any faulty stages of blades.

Once the compressor has been finalized, I will shift my focus to the combustion chamber and digital control of the fuel ratio using Arduinos and sensors sourced from RS Components.

I'm an MEng Biomedical Engineering Student at the University of Glasgow. My academic interests are wide-ranging from prosthetics to 3D printing and electronics to microfluidics. Currently, I'm conducting my Master's Thesis project at ETH Zurich - Department of Biosystems Science and Engineering, working on a microfluidic liver-on-chip platform. I'm also an electronics hobbyist, DIY enthusiast, and debating/MUN lover.
DesignSpark Electrical Logolinkedin