Decision Solutions for the 21st Century
Gas Turbine design - tackling uncertainty at the Design Phase for a Safety Critical Component
The gas turbine engine internal air system provides cooling and sealing air to a series of critical subsystems and components such as high pressure gas turbine blades, as well as controlling the thrust load on the turbine and compressor spool assembly. Many potential variations for the internal air system are possible, depending on the requirement, expertise and command of intellectual property. Some subsystems, such as rim seals, pre-swirl systems, and rotating cavities have been the subject of extensive development and analysis leading to robust design solutions. Nevertheless there remains scope for further consideration of the overall system design, and decision analysis tools such as morphological analysis can be applied to identify early stage design issues for the internal air system. This form of strategic options analysis and supporting processes, provides an effective means for tackling issues where there is uncertainty, as is the case with many design scenarios and where some specific parameters and information is not available until later in the design phase, after the key geometry has been defined.
Case 2: Motorcycle helmet design problem - What are the main factors that contribute to better protection of the head during a motorcycle accident?
During a series of both off-line and group facilitated sessions a specialist team of engineers established that the problem was broken down into two core components – Impact Conditions and Helmet Design. In turn the main focus question was divided into two supplementary questions, each relating to one of the two core components. Thus for Impact conditions the team agreed that: “Given the impact conditions what type of head injury is likely to occur?’ and for Helmet Design: “What are the material design components of the helmet that mitigate/prevent this type of head injury?”
The core objective was to determine what helmet designs were preferable subject to different types of impact in an accident. It was seen early on that it was unlikely that one discrete solution would suffice to offer protection under all conditions, and thus by breaking down the problem into a series of main variables the scale of the problem was addressed.
Crucial to the exercise was identification of those scenario configurations which were internally consistent and which in turn would help the research team in not pursuing unworkable design paths that might manifest themselves further into the design process.
The problem space relating to “Impact Conditions” is shown below as item1.
The problem space relating to the Helmet Design was determined as item 2.
This two tier representation, and based on facilitated input by the team of experts, established that there were 5 core variables in each of the two matrices with an additional LINKING parameter attached to each matrix and which addressed a specific issue in the focus question: the type of head injury to be mitigated according to both impact and design of the helmet.
Item 1 identified an overall problem space of 13824 unique configurations whilst item 2 identified an additional 5760 configurations. If we had combined all 11 different parameters (5+5+1) then the combined number of configurations in the model would rise to over 1.37 million configurations.
The process the research team was able to adopt, was to reduce initially the Impact Conditions problem space to a much smaller solution space, using Fibonacci software. The solution space identified a much reduced set of configurations or scenarios where different forms of Head Injury were the output.
A second exercise was run for the Helmet Design problem space which also reduced the problem space to a small number of internally consistent configurations.
As both sequences had a common parameter – Type of Head Injury – the team was then able to match viable outcomes from each of the two components - Impact Condition and Helmet Design – using the software scenario list to establish those configurations which could work.
The main outcome of the project was to support the design team in structuring an engineering problem by addressing both technical and behavioural variables.