|INSIGHT: Overcoming the Challenges of Centrifugal Compressor Design
High performance centrifugal compressor design remains a remains a great challenge for designers. In this month's issue of The Flow, we sit down with Bob Ni to explore how advances are being made to help designers overcome obstacles to success. Prior to founding ADS, Bob spent nearly 30 years at Pratt & Whitney leading turbomachinery CFD in support of compressor and turbine design.
FLOW: How are companies conducting centrifugal compressor design today?
BOB: Generally speaking, we'd observe that most are continuing to evolve well-understood designs based on a heavy diet of experience and rig testing, with some basic 2-D and 3-D steady analysis for general directional guidance. It's an expensive and time consuming approach, but unfortunately the path of least resistance because the trust in the CFD to properly capture the flow physics still isn't there.
FLOW: Why isn't the trust there?
BOB: I think the bottom line is that the CFD predictions don't corroborate well with the rig data. Candidly, we're not surprised. As designs have become increasingly loaded over the years in an effort to cut weight and improve fuel efficiency, it's become clear that 3-D steady analysis is no longer sufficient to approximate what is inherently unsteady flow behavior inside a centrifugal compressor. This leads to gaps between predicted and measured performance, and worse yet, unforeseen field failures due to high cycle fatigue.
FLOW: What's being done to overcome these challenges?
BOB: We're seeing the more forward-thinking companies incorporate unsteady analysis into the design cycle. By bringing time accurate studies to the front of the development process, they're better able to anticipate and counteract the interaction effects that negatively impact performance and durability.
FLOW: Why hasn't this been done before?
BOB: Unsteady analysis has historically been very computationally intense, rendering it impractical for most design cycles. Fortunately, this is changing quickly with the emergence of solvers like Code Leo, which are delivering excellent predictive accuracy and order-of-magnitude speedups in turnaround time compared to conventional CFD codes.
FLOW: What can be gained?
BOB: At the end of the day, it allows organizations to design with greater confidence. Here at ADS our clients benefit from speedups delivered by Code Leo for both steady and unsteady analysis; these speedups buy time to conduct more thorough studies to improve performance and mitigate the aero loading effects that lead to high cycle fatigue. And by conducting these analyses before committing to hardware, there's a very real opportunity to save a lot of time and money in redesign.
FLOW: Isn't unsteady analysis out of reach for smaller design shops with limited computing resources?
BOB: Absolutely not. There are multiple ways for smaller design shops to exploit unsteady analysis without the need for large internal clusters. For example, you can run unsteady on just a sector of your stage configuration instead of the full wheel. This is straightforward when the blades in each row share a common multiple. For example, a full wheel 24 blade: 12 vane configuration could be reduced to a 30 degree sector (2 blades : 1 vane), thereby reducing computing needs from 36 cores to 3 if 1 core were assigned to each passage. In the event the blade : vane counts don't share a common multiple--24:19 for example--it is also possible to scale a row up/down (e.g. 24:20 or 24:18) and then conduct analysis on a sector (6:5 or 4:3).
Another alternative is to take advantage of cloud computing from vendors like Amazon. The beauty of this approach is that designers can rent what is essentially limitless computing capacity by the CPU-hour at very reasonable rates. Here at ADS we use Amazon EC2 extensively, which we touched on in the May newsletter. Suffice it to say that it's been tremendously empowering for us and has the potential to be a real game changer for small-to mid-sized design shops.
FLOW: Thanks, Bob.
BOB: My pleasure.
CASE STUDY: Reduction of Unsteady Forcing in a Vaned, Contra-Rotating Turbine High pressure turbine blades frequently exhibit unsteadiness due to strong interaction with downstream vanes. In particular, blade unsteadiness in the presence of a downstream vane consistent with contra-rotation may be characterized by strong interaction at the first harmonic of downstream vane passing. In this case study, Dr. John Clark from the United States Air Force Research Laboratory investigates ways to reduce such blade-vane interactions in a 1.5 stage transonic turbine rig. <more>
TECHTIPS: Faster Speedline and Compressor Map Generation with Code Leo
Speedlines and compressor maps can be generated more efficiently in Code Leo by varying back presure against a partially converged restart file. <more>
Welcome to The Flow
Welcome to The Flow, a newsletter for monthly insights on turbomachinery CFD published by AeroDynamic Solutions, Inc.
Each month we'll spotlight a topic of interest, discuss a case study and/or provide useful pointers about how to get the most out of the ADS CFD system.
You are receiving this email because you or someone else you know thought you would be interested. To unsubscribe please click here. We value your privacy.