5/13/2019 | 8 MINUTE READ

Use 3D CFD Analysis to Achieve the Right Conformal Cooling Design

Originally titled 'Get the Full Story on Conformal Cooling'
Facebook Share Icon LinkedIn Share Icon Twitter Share Icon Share by EMail icon Print Icon

Moldmakers must look beyond heat sink simulation and use 3D CFD analysis to achieve the right conformal cooling design.

Share

Facebook Share Icon LinkedIn Share Icon Twitter Share Icon Share by EMail icon Print Icon

Today, there are still conformal cooling skeptics, and because of incomplete analysis, they should be. Simulation is a great validation tool, but that is all it is. It is a powerful tool, but results are still based on the input of the users. Of course, there are optimization tools and modules within the software, but the saying “garbage in garbage out” applies. Two analysts running a plastic injection simulation can give totally different results if the inputs are different. For conformal cooling, specifically, wrong assumptions and dismissing important factors, such as the ability to achieve the necessary flow rate to ensure turbulent flow, can result in a bad investment.  

 

Simulation Vs. Analysis

Conformal cooling can help minimize cycle time and achieve dimensional tolerances. However, it will only work as well as the simulation if the designer can maintain the necessary flow rate through the conformal cooling circuit. A designer running a heat sink simulation (cooling channel defined only as a heat sink source) on complex designs will not tell the full story of what may occur in the mold. Let’s look at some results in Figure 1, which are from a computational fluid dynamics (CFD) analysis showing Reynolds number at a flow rate of approximately 2.0 gpm. 

Some regions show a Reynolds number above 4200, indicating turbulent flow, while other regions show a Reynolds number below 4200, indicating that the flow in certain regions may not be turbulent necessarily. This variation will vary cooling efficiency throughout the circuit. Velocity has a direct correlation to the Reynolds number, so if a designer identifies the low/no flow regions he or she can optimize the cooling channel design. Simulation can also display velocity results that indicate low/no flow regions that are doing less work and potentially stagnant areas. Keep in mind that fluids take the path of least resistance, so instead of taking random paths and flowing around all of the holes, the coolant will flow straight from the inlet side to the outlet side.

Designers can use such data to optimize the design to minimize low/no flow regions (shown in dark blue between the holes). For example, designers can try to offset the holes to divert the water through the regions with a low Reynolds number, which indicates low/no flow regions. Despite a designer offsetting the holes, a path of least resistance may remain. However, offsetting the holes can still help to minimize the no/low flow regions, maximizing turbulent flow throughout with an achievable flow rate. 

Heat sink analysis alone also does not work well for conformal cooling channel analysis because conformal cooling typically consists of small-diameter circuits with long flow lengths that conform to the geometry of the parts. The small diameter or thickness of the design and longer flow length circuits require higher pressures to push sufficient flow rate through the circuit effectively.

For conformal cooling, wrong assumptions and dismissing important factors, such as the ability to achieve the necessary flow rate to ensure turbulent flow, can result in a bad investment.  

 

Pressure Requirements

If Reynolds number and velocity are dependent on flow rate, identifying the pressure required to achieve a certain flow rate is essential to determine whether or not achieving the required flow rate is realistic. Figure 2 shows coolant pressure results from the same model at the same flow rate of 2.0 gpm. 

A designer must understand pressure requirements and the minimum flow rate to achieve turbulent flow in the design where they intend to have good heat transfer because fluids take the path of least resistance. Figure 1 indicates that water flows in areas where the Reynolds number is between 4,000 and 10,000+ and that turbulent flow (Re # > 4200) is likely present in most regions. However, high pressure is required to flow 2.0 gpm through the circuits to achieve these values.    

Thermolators on the market today can generate up to approximately 100 PSI, but 65 PSI might not be achievable because fluids prefer the path of least resistance or areas with less resistant circuits, such as manifold plates and core and cavity plates. If the flow rate falls below 2.0 gpm, the Reynolds number will decrease, potentially yielding laminar flow through the circuit, which is not efficient. Also, a designer simply increasing the flow rate to the manifold will not result in a linear distribution of fluid through all circuits.

Simulated results and actual results may differ if a designer does not hook up the conformal cooling circuits to their thermolator or if they don’t use flow restrictors to divert water to specific lines. If the designer cannot achieve the necessary flow rate to maintain turbulent flow through a conformal cooling circuit, they can drill conventional circuits and achieve equivalent or better results. 

A designer must understand pressure requirements and the minimum flow rate to achieve turbulent flow in the design where they intend to have good heat transfer because fluids take the path of least resistance.

 

Meshing Matters

A designer can achieve a more accurate and reliable analysis of conformal cooling designs with accurate CFD analysis. However, not all CFDs are the same. The two main meshing methods for cooling channels in injection simulation software are beam elements and 3D elements. Beam elements are generally used for long, thin geometries where one dimension is much larger than the other two. The calculation is approximated from the beginning node to the end node. 3D elements tend to provide a complete geometric representation of a design and can capture the physical phenomena occurring during calculations at the nodes of each mesh element.

Beam elements can determine the Reynolds number, temperature rise/fall throughout the mold, and flow distribution, but they are not nearly as accurate as 3D mesh elements when predicting pressure drop. Beam elements also have a hard time capturing true complex geometries and require modeling as relative volume due to the nature of the beam element. However, beam elements do yield faster simulation times compared to 3D mesh since there are a lot fewer calculations occurring throughout the system. 

Figure 3 compares the pressure drop results between beam element and 3D mesh simulation methods. Pressure drop results indicate the head pressure required to push the indicated flow rate through the cooling circuits. Although the pressure requirement ratio from one circuit to another is similar, the beam element indicates 25 to 50 percent less pressure required for the complex circuits and the potential to exacerbate the pressure with a complex conformal cooling design. Pressure drop results alone can lead to false thermolator requirements, which will yield insufficient flow through circuits that are more restrictive — for example, conformal cooling circuits.

A designer can achieve a more accurate and reliable analysis of conformal cooling designs with accurate CFD analysis. However, not all CFDs are the same. The two main meshing methods for cooling channels in injection simulation software are beam elements and 3D elements.

 

Proving It Out

To determine if 3D mesh elements capture more precise pressure drop prediction than beam elements, Kevin Klotz from MGS Mfg. Group conducted a test to validate pressure predictions (see Figure 4). 

Klotz used one DMLS-conformal-cooled insert DMLS, a 3D-printed set of bubblers, clear tubing with different length and diameters, pressure gauges and flow gauges connected to a water transfer pump and a water tank. The apparatus was modeled in Creo and recorded to compare with simulation. The same model was sent to Dr. Jeet Sengupta of Hoerbiger Corporation of America to run a CFD analysis via Ansys CFX and to simulate 3D boundary layer meshing (BLM) using Moldex3D. 

Both software solutions use the recorded flow rate values through each circuit and the combined flow rate value of 6.5 gpm as the input from the pump to record pressures at the inlet, exit and pressure gauge locations. Figure 5 shows the results. The designer using Moldex3D modeled the inlet and outlet hoses straight instead of curved because this software requires inlets and outlets of the fluid source to be perpendicular to the mold face. 

Pressure results from both software tools are very close to the recorded values from the physical apparatus. Table 1 compares the recorded values to the simulated pressures. For example, the largest difference from Ansys CFX to the recorded values is about 5.0 psi and 2.5 psi for Moldex3D using BLM mesh. Both results are well within acceptable accuracy for the intended purposes, considering the tolerance in pressure gauge reading and flow rate measurements from an inline flow gauge.

 

Lessons Learned

Conformal cooling can work as intended if it is properly designed, which requires a thorough simulation using CFD analysis to understand the characteristics of the fluid flow. For example, if a 3D-printed insert branches into two or more channels, determining the favored branch can make a large impact on conformal cooling design effectiveness.

When a designer only runs a heat sink analysis, the mold temperature profile may be incorrect if there is inefficient flow through a circuit. For example, if there is insufficient pressure to maintain good flow through a circuit, the coolant can potentially absorb too much heat, causing too large of a temperature difference between the inlet and outlet. The coolant can also become less effective at removing heat from the mold as the water travels through the cooling circuit.  A heat sink analysis can underpredict the required pressure to achieve a flow rate for the turbulent flow of complex cooling designs such as conformal cooling. However, a designer can precisely capture fluid flow characteristics using 3D mesh elements, such as momentum change for accurate pressure predictions.

The effectiveness of a complex cooling design is highly dependent on the ability to reach turbulent flow throughout the entirety of the system. Simulation has come a long way to attain the accuracy and performance necessary to weigh the benefits of the more complex designs, and it is up to the designer to choose the most efficient computation method to achieve the right conformal cooling design.

 

About the Author

 

Jay Vang

Jay Vang is a customer success engineer for Moldex3D Northern America.

For More Information

 

Moldex3D Northern America

248-987-7987 ext. 992 / jayvang@us.moldex3d.com / epsflotek.com/ ​​​

 

Ansys Inc.

ansys.com

 

MGS Mfg Group

mgsmfg.com


RELATED CONTENT

Resources