I think we can all agree that designing turbomachinery is hard. There are just so many moving parts (pun intended) in the design process, and they are all interconnected. When you change the blade shape, it changes the aerodynamics, and could impact manufacturability. Everything you change has a cascading effect across many different areas, because all of the areas are linked; just like a Rubik's® cube! Only, in turbomachinery design, you are not always trying to get all of the sides to be one color. Heck, even a 3-year old can do that.
Imagine that each side of the cube represents various aspects of the design process (geometry, CFD, FEA, manufacturability, etc.) and there are different patterns of colored squares on each side that represent the specs you are hoping to meet. Every time you adjust one aspect, say, change the blade geometry, it changes the other sides too. And since turbomachinery design is, by its very nature, an iterative process, it gets tricky fast.
Consider designing a single-stage pump to move wastewater. The design specs are probably fairly simple, analogous to a 3x3 Rubik's cube. Sure, there are challenges, but it is pretty straightforward. Now, you are working on a radial compressor for a chiller that uses one of the new environmentally-friendly refrigerants. Its getting harder, now we are talking about a 4x4 cube. What about a turbocharger? Simple, right? They have been around for over a hundred years, and there are MILLIONS of them produced every year. Are we back to a 3x3? Given the broad operating range and relentless drive towards efficiency, I'd say a 5x5. What if your next job is a multistage gas turbine used in power generation? All those stages and operating temperatures add a lot more complexity. We are talking at least an 8x8. Then there are turbines used in aerospace and aviation. As you might imagine. whenever there is a risk to life and limb, the complexity skyrockets to at least a 20x20.
Just like solving the Rubik's cube, you can design turbomachinery by manual trial and error, or you can use algorithms to help speed up the process. For Rubik's Cubes, it is a set of "rotation sequences"; in turbomachinery design, it's software. Available software ranges from "general purpose" Computer-aided Engineering (CAE) and Computer-Aided Manufacturing (CAM) software that can be used to design a plethora of things, to software designed specifically for turbomachinery.
It gets harder and harder to design things "manually" or with simple tools, as the complexity increases. Sure, it can be done, but at what cost? The phrase "time is money and money is time" definitely applies to designing pumps, compressors, turbines, and turbochargers. For example, take mapping the static pressures, relative total temperatures, and the heat-transfer coefficients to all of the flow path surfaces in the finite element model (FEM). You can do this manually using general software, but it will be a long, slow process, prone to errors from moving data from program to program. And you have to do it each time you change your design! How many iterations are you going to do?
Another option is to use an automated process. Like this machine that can solve the Rubik's Cube in under 1 second, sophisticated, specialized software can make the design process much faster. The aero-load mapping feature in Pushbutton FEA™ can, in two clicks, do what it takes hours or even days to do manually, using multiple software packages. Pushbutton FEA automates the link between the CFD results (pressure, temperature, heat transfer coefficient) to the structural FEM for both primary and secondary flow paths. With this tool, you can reduce overall design process time, or perform numerous iterations. Most users opt for a balance of the two.