Simulation tools allow engineers to explore, test, validate, and optimize product ideas and designs through the use of computational fluid dynamics (CFD), heat transfer, and structural analysis. The advanced physics solvers available to designers and engineers empower them to study & optimize pressure drop and force behavior, evaluate fluid flow patterns, minimize (or even leverage) cavitation effects, analyze vibration impacts, and more. By deploying simulation early in the product design process, turbomachinery and pump engineers can ensure that they meet and in most cases, exceed their design validation requirements.
By Kanchan Garg, Product Manager, SimScale
Why Simulate Pumps and Motors
Modeling provides detailed insights into performance characteristics, allowing engineers to identify design strengths and weaknesses in a virtual prototype, avoiding costly physical testing & prototype rework. The resultant time & cost savings from simulating early and often allows engineers to explore more of their ideas, sparking innovation which leads to competitive insights. For example, a typical workflow is to import a Computer-Aided Design (CAD) model of a pump, select what analysis type to perform (fluid, structural, thermal), set up the simulation parameters and boundary conditions including material selection, discretize the CAD domain (Meshing), and simulate and analyze the results. Legacy simulation tools limited to desktop hardware have been a limitation to the wide scale and economical adoption of simulation tools for engineers in all industries. A new generation of cloud-native simulation tools accessed via a web browser, with faster algorithms and simpler user interfaces, are enabling the use of simulation across teams and entire organizations and earlier in the design process.
Advances in Turbomachinery Applications
Turbomachinery products and components need a special set of simulation tools due to their unique properties. Modern cloud-native tools now have robust CAD handling capabilities that let engineers spend less time on making models’ simulation-ready, and more time on analysis. Automated binary tree based meshers can do this without any input from the engineer, who can spend more time evaluating subsonic and supersonic flow behavior, turbulence effects, cavitation, and multiphase flows. The accuracy of these solvers is close to 2% as compared with test data, and a designer can calculate an entire pump curve, by simultaneously running multiple simulations in the cloud, in 15 minutes. This is possible using input parameterization for fast design prototyping and CAD associativity for easy geometry variation.
The toolset is encompassing and includes multiphysics analyses such as:
- Static and dynamic analysis for peak structural stress, deformation, rotor/shaft strength analysis,
- Vibration and harmonic analysis to reduce noise and longterm fatigue,
- Thermo-structural shock analysis,
- Transient thermal stresses during pump operation.
Designers of pumps can now generate an entire pump curve for a specific blade geometry and calculate performance curves for multiple pump designs in parallel. Making use of application programming interfaces (API) extends the use of simulation tools to include third-party optimization and design of experiment (DOE) tools.
Case Study
Heating circulator pumps are commonly used in heating ventilation and air conditioning (HVAC) applications. As per the ErP ecodesign directive of the EU, these pumps must adhere to stringent requirements, expressed by the Energy Efficiency Index (EEI), to be sold in various markets. Similar standards exist in other parts of the world. EEI is calculated by comparing the average power consumption at four weighted flow rates against a predefined reference power input and is determined by the hydraulic design of the pump, the motor, and the electronic controller. These system components vary in their design and selection until very late in the product development cycle, meaning that the hydraulic design needs to be done fast (at this late stage) so it fits the power curves of the selected components. In this study, the hydraulic part of the pump (impeller) was developed using an innovative automated workflow that is easy to replicate.
Design Parameters
Its design was based on computing a large set of impeller design variants for multiple operating points using high- fidelity CFD simulation. 14 pump impeller geometry parameters were varied, each at three pump flow rates (operating points), resulting in 300 design iterations. The generated performance data was stored in a surrogate model (reduced order model) that can then be interrogated to enable rapid impeller selection in the final stages of the product development process.
The chosen design variants were simulated using the cloud-native SimScale platform to generate performance curves for each one and fed into CAESES, a shape optimization tool that can take robust geometry models into any simulation-driven optimization loop. The ultimate goal is to create a surrogate model for the pump impeller, which is prepared in CAESES using performance insights from the software. The surrogate model is created and further optimized and the two tools communicate with each other using the API. At the time of production, based on the rest of the system components, the best impeller design can be chosen and simply inserted into the system before production.
Summary statistics for the simulations performed on the 300 design variations:
- The actual simulation runtime for the 300 design variations was 42.4 hours and benefitted from parallel simulations in the cloud. If these had been run serially using legacy CFD tools, the total cumulative runtime would have been 592.6 hours (25 days).
- A total of 900 simulations were run (300 design variants at three operating points, each) consuming 3,000 core hours of cloud computing at a cost of $300 (For this particular design of experiment only).
Pump Motor Case
Motors in pumps are critical components that require similar levels of diligence at the early design stages. They are subject to various types of loads that might interfere with performance and lifetime. Motor vibrations can cause resonance and harmonic distortion leading to displacement in the structure which can damage the motor. The ‘modes’ at which the motor is susceptible to excessive vibration should then be modeled using modal analysis to evaluate the limits of physical displacement and/or even deformation. Additionally, heat transfer and flow analysis are required for thermal management, especially where additional cooling is required. Pressureflow analyses help engineers to determine the optimal flow rates and coolant types. Torque stresses help to select the right materials and size of shafts for best performance.
A multiphysics analysis of a common electrical motor used in larger pumps is illustrated in Table 1, and shows different physics including:
- The CAD model of the motor as built (1a)
- Fluid flow (CFD) through the helical cooling jacket around the motor stator (1b)
- Conjugate heat transfer (CHT) from the fluid in the cooling jacket to surrounding heat-exchanged fins & motor structure (1b)
- Vibration and Harmonic/Modal simulation of the rotor (1c)
- Torque stress on the motor shaft (1d)
- Pressure drop across the coolant channel (1e)
Summary
Using simulation to optimize the design and performance of pumps is essential for innovating high-performing pumps that last. With advances in simulation technology and the associated infrastructure of access, storage, computing power, and collaboration all in the cloud, accessed via a web browser, engineers can leverage these tools more easily than ever before.
ABOUT THE AUTHOR
Kanchan Garg is an aerospace engineer and computational fluid dynamicist by training. She now drives the development and promotion of SimScale’s technology in the rotating machinery industry.