Joe Boothroyd, Analysis Engineer at Gilkes, shares insights about his dissertation project.
We recently celebrated the graduation of three Gilkes engineers. To showcase their work, we are starting a series of blogs focusing on their projects. In the initial post, Joe Boothroyd discusses his dissertation project and its practical relevance to his current position as an Analysis Engineer. The blog details the validation process of computational fluid dynamics for predicting performance in reverse engineering Francis turbines. Joe explains why he chose this subject, what his objective was, and how it aligned with his current role at Gilkes.
Computational Fluid Dynamics Validation of Reverse Engineering Performance Prediction for Francis Turbines
With increasing focus on hydro refurbishment projects, Gilkes is occasionally having to rely on reverse engineering the existing runner geometry when limited or no drawings are available. I’ve been involved on a couple of these projects using 3D scanning to create a CAD model for CNC manufacture for reverse engineering of Francis runners. This project gave me the chance to combine two of my specialist areas (CFD and scanning) to better understand the accuracy of this process.
3D scanning process
Understanding the accuracy of the reverse engineering process
The aim of this project was to understand how accurate the reverse engineering process of Francis runners can be and how effective computational fluid dynamics (CFD) is for understanding any hydraulic performance loss throughout the process. I.e. Can Gilkes receive an old/ damaged runner that they do not have any design information for, 3D scan and CAD model the existing runner and then produce a like for like new runner that has not altered in hydraulic performance.
The Loch Mannoch runner refurbishment work was the basis for this project, using the following process:
- Gilkes had access to design information for the original Loch Mannoch runner, from which I first created a CAD model. This would produce the ideal runner design.
- Gilkes then used this CAD model to have a new runner CNC manufactured from a solid billet.
- I then scanned and reverse engineered the new manufactured runner, again creating a CAD model. I then had 2 CAD models.
- A perfect, from drawings CAD model
- A reverse engineered CAD model from the physically manufactured runner
- I produced a CFD model for both runners and compared the efficiency to understand if there were any geometry deviations throughout the reverse engineering process that effected hydraulic performance.
- I verified the accuracy of the CFD model using test data for the Loch Mannoch runner design.
The results
The outcome showed that Gilkes’ reverse engineering method alone had an accuracy loss of 0.36% when accounting for geometry differences between the reverse engineered runner and the original runner. The outcome of this project gave me confidence that the reverse engineering work I have been carrying out is accurate and there is now a verified piece of work to prove this. The Gilkes engineering team have since gone on to use this method for a number of hydropower modernisation projects.
The project developed my learning by filling the gaps of knowledge on the reverse engineering subject, that could only be gained from a practical project. It allowed me to refine some areas of the process for use in Gilkes’ future reverse engineering projects, while at the same time, gaining another skill to support my future IMechE engineering chartership application.
Joe Boothroyd, Analysis Engineer 29/01/2025
Diagram demonstrating the scan to manufacture process
Scanning of a large Francis runner