Digital twin tackles design challenges

Figure 5. Flow streamlines in pump discharge colored by velocity magnitude at 359 (left) and 1110 (right) Gallons Per Minute (GPM)..

Turbomachinery designers are now using advanced design methodologies such as Computer-Aided Engineering (CAE) tools for virtual prototyping, optimization and troubleshooting. Simulation tools enable designers to accurately predict performance earlier in the design cycle, to analyze multiple designs, reduce reliance on multiple physical prototypes and expensive testing, optimize design for maximum performance and shorten design time and cost.

MSI used Siemens’ STAR-CCM+software, which includes features to help product development organizations enhance and accelerate their ability to digitally simulate and understand how a product will perform in the real world using the digital twin, which is a precise virtual model of a product’s physical and performance characteristics.

Industrial water pumps in a factory.

Simulating water pumps

Water pumps are widely used in fossil fuel, nuclear and hydro power plants, as well as in other industries like chemical processing, automotive cooling, oil & gas and waste management. The major design challenges of water pumps include the unsteadiness of the flow, cavitation and rotating machinery.

STAR-CCM+ offers features that can successfully tackle these difficult physics while offering a streamlined, robust numerical design process. Geometry generation, meshing, solving, post-processing and optimization can all be done in an efficient workflow from within STAR-CCM+. Features like the unsteady flow solver, unsteady cavitation model, Rigid Body Motion (RBM) for rotating domains, automated unstructured meshing for complex geometry and parallel processing capability all offer the ability to explore multiple water pump designs in an economical manner. STAR-CCM+ allows designers to easily predict pump performance at design and off-design points and avoid the damaging effects of cavitation and erosion.

Vertical flood control pump

This example focuses on a large axial pump for the city of New Orleans, designed to provide durable performance under severe weather conditions. These new vertical flood control pumps were intended to remove excess water from the city during flooding and MSI was involved in analyzing the pump design for resonance and vibration issues resulting from an unsteady, forced response. The pump had a large inlet, an impeller with a wheel diameter of over one meter, a vane diffuser, an internal shaft, a discharge vane propped by guide vanes and an exit siphon. The complex geometry of the pump was imported into STAR-CCM+ and discretized using the automated polyhedral and prism layer meshing capability. The final computational mesh had four million volumetric cells and captured the complex geometry with multiple channels and flow passages accurately.

Figure 1 shows the complex geometry and computational mesh of the flow path.

Figure 1. Vertical Pump geometry (left) and computational mesh along centre plane (right).

Figure 2. Velocity contours at centre plane shown as Line Integral Convolution (LIC) plot.

The segregated flow solver in STAR-CCM+ was chosen with the second order convection scheme and realizable k-epsilon turbulence model. Water was chosen as the working fluid at standard temperature and pressure. A mass flow boundary condition was specified at the inlet, static pressure at the exit and rotating speed on the domain enclosing the impeller was specified as well. The transient simulation was run to complete one full impeller rotation with 20 inner iterations per time step. The simulation was run for several complete rotations until the residual monitors were settled and variable monitors for pressure, torque and mass flow showed cyclical behaviour.

Figure 2 shows contours of velocity at the centerline plane of the pump. The image shows vortical flow through the discharge region created by the shaft. STAR-CCM+ has been well validated at MSI and hence, pressure traces from the simulations were used for frequency analysis with LabVIEW instead of experimental values. The pressure at every interface in the domain was processed in frequency domain in LabVIEW using Fast Fourier Transform (FFT) for spectral analysis. The frequencies of oscillation in the flow path were captured in this analysis including vane pass frequency, impeller frequency and two times vane pass (Figure 3). The captured frequencies were analyzed in amodal analysis software and no vibration/resonance issues were found in the pump design.

Figure 3: Relevant impeller frequencies from STAR-CCM+ at impeller & diffuser inlets.

Multistage pump

This example details the simulation work done by MSI on a multistage pump for a customer interested in analyzing the performance of the pump design with an unusual outlet. The centrifugal pump with multiple stages was designed to provide large amounts of Total Developed Head (TDH). In addition to complex transient flows, rotating machinery and unsteady forced response, the pump also involved complex secondary flows and it was paramount that the simulation tool captured all these relevant physics in the analysis. The pump simulation involved the flow from the inlet going through two stages with two impellers in series and discharging through a volute. A thrust balance device redirected a part of the exit flow back to the inlet. The pump had an extremely complex primary and secondary flow path and was designed to enhance cavitation and hydraulic performance.

The flow domain of the multistage pump included the inlet, discharge, primary and secondary flow passage and interfaces connecting the stationary and rotating regions. Figure 4 shows the geometry and the computational mesh on the pump. A total of eight primary and secondary interfaces connected the stationary and rotating regions in the simulation. STAR-CCM+ was used to solve the flow around the impeller, the cavity regions in front and rear of the shroud for each element, in addition to the flow inside the tubing connecting the inlet to the outlet.

Figure 5 shows the comparison of flow streamlines at a low and high flow rate through the pump. At low flow, the left side of the volute shows different flow characteristics compared to the right with flow being choked off and non-symmetrical on the left side. At high flow, the flow is vortical but more uniform in both sides.  

Figure 4. Geometry (left) and computational mesh (right) of pump in STAR-CCM+

The change in total head levels off as the flow rate moves from high to low value. From the STAR-CCM+ simulations, the non-uniform flow was identified as the cause of performance deterioration at low flow rates. This analysis and observation were very useful for the customer in redesigning the pump for producing proper head at lower flow rates.


MSI has successfully incorporated STAR-CCM+ into their process chain as a design and troubleshooting tool with excellent results. The advanced, accurate flow solver in STAR-CCM+ offers MSI a streamlined, cost-efficient engineering process to solve extremely complex problems in hydraulic turbomachinery. The examples described here clearly show the benefits of numerical simulation using STAR-CCM+ for turbomachinery design and analysis in an industrial setting.

Figure 5. Flow streamlines in pump discharge colored by velocity magnitude at 359 (left) and 1110 (right) Gallons Per Minute (GPM).


Stephen Ferguson, marketing director, Siemens PLM Software


Edward Bennett & Artem Ivashchenko Mechanical Solutions, Inc. 11 Apollo Drive Whippany, NJ 07981-1423 USA Tel: (001) 973-326-9920 Email: Website: