With the right setup, this software can open new possibilities for system and pump design improvements.
by Mark R. Anderson (Concepts NREC)
May 1, 2015
Figure 3. Flow field result for the baseline and optimized volute. Note the more uniform velocity distribution after the tongue. (Courtesy of Concepts NREC)Figure 3. Flow field result for the baseline and optimized volute. Note the more uniform velocity distribution after the tongue. (Courtesy of Concepts NREC)

Efficient parameterization of the inputs is a critical factor in the success of an optimization. The case is a good example of how a system-level approach is the most cost-effective.

The design space must be explored as efficiently as possible to ensure that a reasonable number of iterations are needed to actually get to the optimum. Systems using parameters that define a wide variety of coherent shapes with a minimal number of inputs are superior to other methods.

The second study also used CFD, but this time to optimize the impeller. A snapshot of the grid used to analyze the baseline design is shown in Figure 4. The analysis included 14 input variables. Ten different outputs were considered from the solver, either as constraints or included directly in the objective function.

Figure 4. Cross-section of the structured grid example for the pump impeller optimization (Courtesy of Concepts NREC)Figure 4. Cross-section of the structured grid example for the pump impeller optimization (Courtesy of Concepts NREC)

The result of the process yielded a design with 1.2 percent higher efficiency at the design point, with no penalty at the off-design flow rates. Head rise went from 99 to 104 feet at the design point. About 450 CFD solutions were required by the optimizer (see Figure 5).

igure 5. Solution for the pump impeller optimization (Courtesy of Concepts NREC)Figure 5. Solution for the pump impeller optimization (Courtesy of Concepts NREC)

Conclusion

Automated optimization is becoming more widely used in the turbomachinery industry at many levels. While commercial offerings have improved the process significantly, making effective use of optimization still requires expertise on the part of the user.

That said, a properly designed optimization exercise can both streamline the entire design process and discover improved performance from regions of the design space that the user might never have considered previously. Expanded capability in modeling methods and reduced cost in high-powered computing resources both point to the increasing use of advanced optimization methods in the pump industry.

Acknowledgements
Dr. Edward P. Childs of Concepts NREC performed most of the optimization setup and analysis shown in this article and laid much of the ground work for the optimization system shown.

References

  1. Japikse, D., and Platt, M. J., "Optimization in Component Design and Redesign", 10th Int'l Symposium on Transport Phenomena and Dynamics of Rotating Machinery (ISROMAC-10), Honolulu, Hawaii, March 7 – 11, 2004.
  2. Anderson, M. R., "Optimization of Turbomachinery – Validation against Experimental Results," CDK Nové Energo, 2009 – Current Trends in Design and Computation of Turbomachinery, June 2 – 4, 2009.

Pages