# Evolution Potential Hierarchies

| On 08, Mar 2018

Darrell Mann

Although widely established as a means of analyzing the maturity of a system, and for guiding the generation of future evolution direction ideas, there continues to be confusion over how to get the best out of the Evolution Potential process. I thought Iâ€™d use a simple example to illustrate a technique we frequently use within the SI research team to make sure our efforts to get the maximum number of evolution ideas out of using the tool.

Well, when I say â€˜simpleâ€™, and I think back to the time I first had the job of designing centrifugal impellers, it didnâ€™t feel very simple at all. This is back in the mid-1980s, before I knew about TRIZ. It was the pioneering age of computational fluid dynamics (CFD), so in reality, my main â€˜designâ€™ task was designing the software that would allow us to subsequently design better â€“ more efficient â€“ impellers. At the time we were pioneering the use of 3D analysis tools that allowed us for the first time to model what was happening as air was induced to flow through the impeller. The software made two significant evolution jumps: when I first joined the department, impellers were effectively designed using a 1-dimensional calculation. A lot of the calculations were done using calculators and the job was highly tedious. Then we got 2D software, and now I was developing the 3D version. The design of the impeller itself, however, was lagging someway behind â€“ the shape of the vanes effectively being determined, still, from the 1D and 2D calculations.

Hereâ€™s what weâ€™d have drawn if weâ€™d known about TRIZ back in those days:

Figure 1: Evolution Potential Analysis Of Typical Centrifugal Impeller

What it would have told us is that, just like weâ€™d seen with the evolution of our software-based design tools, the impeller geometry was ultimately going to become more three dimensional.

That said, to the lay-person, the above impeller picture already looks very three dimensional. So why are we able to say that it will become â€˜moreâ€™ three dimensional? The trick to the realization â€“ and the scoring of the axes on the Evolution Potential plot â€“ is to look at the overall geometry of the impeller and score according to the lowest point along the relevant trend scale. Thus, although the impeller looks 3D, when we start looking at many of the individual features we see that they are straight lines. Hence, the overall impeller has been evaluated as being at the second stage (â€˜lineâ€™) along the Geometric Evolution trend.

Figure 2 highlights some of the â€˜straight-linesâ€™ we can see on the impeller:

Figure 2: Straight Line Features On Impeller

Each of these features represents an opportunity to design a better impeller: TRIZ tells us that straight lines â€˜wantâ€™ to evolve to become curved: the vane leading edges want to be curved; the trailing edges want to be curved; they donâ€™t want to be perpendicular to the base, etc.

I could use my 3D design software to help me to work out how to curve each feature, but until Iâ€™ve exploited each and every straight-line and made it curved, my Evolution Potential plot for the overall impeller will remain unchanged. Only, in other words, when there are no more straight-lines can I say the Geometric Evolution Trend score for the impeller has seen a jump to the next, â€˜curvedâ€™ stage.

The main reason for adopting this kind of discipline is that as soon as I â€“ or the next designer after me â€“ see a plot in which the â€˜make it curvedâ€™ Evolution Potential has been used up, we will stop looking for further straight-lines to make curved.

The problem with this discipline is that it can be an awful long time before my plot shows any progress.

The solution to this problem is to begin constructing a hierarchy of Evolution Potential plots. The high-level plot for the overall impeller may not change for some time, but if Iâ€™ve constructed a plot for a â€˜vaneâ€™, or, even more detailed, â€˜vane trailing edgeâ€™, then I will have a much more comprehensive means of mapping how much potential I still have available to me.

Figure 3 makes an attempt to show what such a hierarchy might look like for our impeller.

For the eagle-eyed reader, you may notice that at the bottom of the hierarchy, the plot has been drawn to include an evolution jump. One of the first things we realized once we started using our 3D design codes was that it was a really good idea to make the vane trailing-edge root orientation such that the vane was no longer perpendicular to the baseâ€¦

Figure 3: Skeleton Impeller Evolution Potential Hierarchy

â€¦letâ€™s have a closer look at the bottom-of-hierarchy plot and the resulting change in the impeller trailing-edge vane geometryâ€¦

Hopefully the idea is simple. The real challenge is in the discipline required to construct and manage the hierarchy. I know if I was back in charge of impeller evolution strategy, Iâ€™d have a wall full of Evolution Potential plots and would be managing my R&D activities according to all of the untapped potential.