|INSIGHT: The New Normal—4D Turbomachinery Design
3D steady analysis is no longer sufficient to provide the flow insights needed to advance modern turbine and compressor design. Design processes must now address the fourth dimension—time—to tackle the unsteady blade row interaction effects that adversely impact performance and durability. In this month's issue of The Flow, we sit down with Bob Ni to discuss how recent advances have made unsteady analysis practical for use during design and how it has been successfully deployed. Prior to founding ADS, Bob spent nearly 30 years at Pratt & Whitney leading turbomachinery CFD development in support of turbine and compressor design.
FLOW: Bob, unsteady analysis has been around for a long time. Why hasn't it gotten more industry traction?
BOB: Though it's well understood that turbomachinery flows are inherently unsteady, time resolved analyses have historically been too computationally expensive for commercial design. So the industry continues to rely on 3D steady analysis and prototypes as a workaround. For industrial application, unsteady analysis is typically relegated to spot duty, after design, when field problems arise and root cause analysis is required.
FLOW: Why are things changing now?
BOB: I think there are three reasons. First, there's a greater sense of design urgency. The quest for more efficient turbines and compressors has led to smaller, more heavily loaded designs that amplify blade row interaction effects which impact performance and durability. 3D steady approximations are inherently unable to capture these effects and can lead to bad design decisions. Second, a new generation of highly efficient solvers like Code Leo have emerged that can provide order of magnitude level speedups in turnaround time for multi-stage unsteady analysis. Faster turnaround time makes unsteady analysis more practical for commercial design. Third, the advent of cloud computing from vendors like Amazon Web Services have made it possible for designers to conduct large scale unsteady analysis on-demand and on a pay-as-you-go basis. This makes it much more feasible for smaller companies to capitalize on 4D analysis without the need for major infrastructure investment.
FLOW: Interesting. How have you observed unsteady analysis best incorporated into commercial design cycles?
BOB: Designers will continue to leverage meanline analysis to narrow down the design space and 2D/3D steady analysis for blade and row design. Unsteady analysis gets incorporated into the process when optimizing the overall multi-stage design. Normally, once all blade rows have been designed, 3D multi-row steady analysis is conducted in some sort of optimization loop to achieve design objectives. In the new normal, 4D (unsteady) analysis is applied instead. A good example of this can be found in this month's case study, where you'll see how United Technologies Research Center was able to take advantage of this type of process to deliver a significantly improved centrifugal compressor design.
FLOW: Wouldn't the incorporation of unsteady analysis lengthen total design time?
BOB: It depends on how you look at it. Looking purely at initial design time, probably. But looking at the overall design process, probably not since a better job is being done of anticipating interaction effects that lead to performance degradation and possibly high cycle fatigue. So the extended initial design time is very likely recovered later in the process where adverse impacts to schedule and cost are greatly amplified. We've seen cases where efficiency predictions between steady and usnteady calculations varied as much as two percentage points.
I'd also add that the impact to design time depends heavily on the speed gains you can achieve relative to your existing CFD solver. For example, here at ADS our clients tell us that Code Leo delivers 2-4x turnaround time improvement for steady analysis compared with commercial and government codes, and 5-30x speedups for unsteady. So it's possible to envision scenarios where initial design time may even been shortened a bit.
FLOW: Many designers tend to choose blade row counts that avoid resonance, which makes it difficult to reduce the analysis domain to anything but full wheel. Isn't this a problem?
BOB: Certainly it's a consideration; as always, you have to trade off time vs. accuracy. Here at ADS we've tackled this issue in two ways. First, for those wanting the highest fidelity results, we've dramatically improved the turnaround time for large scale unsteady analysis and enabled it to be run locally or through Amazon Web Services. Second, for situations where scaled approximations are appropriate, we can enable sector analysis much more easily through an implicit scaling technique we discussed in last month's newsletter. So there are good options, even for those with unusual blade counts.
FLOW: In what situations do you think 4D analysis is a must-have for design?
BOB: Candidly, if you're struggling to innovate with respect to performance and duty cycle, you're going to want 4D analysis to reveal the flow insights necessary to advance your designs. We think this is particularly acute in compact, high work, high efficiency turbines and more generally in centrifugal compressor and radial turbine design, where flows are extremely complex.
FLOW: How do you suggest getting started with 4D analysis?
BOB: The best way to start is to take a representative design where you have both 3D steady predictions and experimental data. Conduct unsteady analysis on this configuration using your existing CFD as well as one of the next generation solvers like Code Leo and assess three things: (1) predictive accuracy; (2) turnaround time; (3) the ability of the code to properly reveal flow insights contributing to performance degradation. Your findings should give you a good feel for whether or not 4D analysis can make a difference.
FLOW: Thanks Bob.
BOB: My pleasure.
CASE STUDY: Design of a High Efficiency Compact Centrifugal Compressor for Rotorcraft Applications In this recent paper from the American Helicopter Society 67th Annual Forum, Beth Lurie from United Technologies Research Center describes the design of a compact, high efficiency centrifugal compressor stage for rotorcraft applications. A CFD-based design methodology that included unsteady calculations of impeller-diffuser interactions was used to assess over 50 configurations at over 400 operating points. <more>
TECHTIPS: Configuring an Unsteady Simulation with Implicit Scaling
Blade row counts are commonly selected in turbomachinery design to avoid periodicity. Learn how to take advantage of implicit scaling in ADS CFD to avoid the manual overhead of explicit scaling and the longer turnaround time of full wheel analysis. <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.