Three challenges in laser powder bed fusion and how to overcome them with simulation

More lasers mean more power, higher build rates and lower costs per part. At least that is the main expectation for the new generation of laser powder bed fusion (LPBF) machines with two, three, four, or even twelve lasers. But the massive energy increase does have unexpected downsides.

LPBF is an additive manufacturing (AM) process in which a laser beam is applied to a metallic powder to build a part layer by layer. To boost the machines’ productivity, printer producers started adding more lasers, which has, in turn, increased the risk of overheating of components, shrinkage, and variations in microstructure and mechanical material properties. Sometimes, even if a part looks right on visual inspection, in actual application it cracks or fails. 

Using a four-laser machine can result in significant overheating of the maraging steel as is visible from the discolouration.

To take the guesswork out of LPBF, manufacturers can digitally simulate the whole process. A thermal simulation allows us to investigate the influences of additional lasers and to estimate their impact on part quality. 

Today thermal simulations strongly focus on single parts, investigating whether support structures or the geometry itself are sufficient to transport the heat to the baseplate. However, to achieve accurate predictions for the temperature of a single part, you need to know the build rate of the entire job. Even when there isn’t any transfer of heat from one part to another, it’s still crucial to consider the real build setup. 

Our simulation software, Amphyon, takes a two-stage approach to achieve an accurate simulation of the LPBF process. Amphyon first calculates build rates based on parameters like the number of lasers, laser power, scan speed, recoating time, etc. Then, it predicts the absolute temperature with a thermal finite element analysis. In this case, we don’t calculate peaks above melting temperatures as that isn’t relevant. Instead, what we need to know is the ambient temperature of the part before the scanning of the next layer starts. If a layer results in too high a temperature, Amphyon can calculate the time necessary to cool down and, in the end, save printing time.

To show you how simulation allows us to avoid overheating, let’s look at three production examples based on real situations engineers are experiencing in LPBF.

Example 1: One laser machine vs four laser machines

You may think the only difference between using a single-laser machine and a four-laser machine is more power, but a simulation reveals there’s more to the process than meets the eye.

Our thermal simulation of the manufacturing of a mould from maraging steel shows that adding three more lasers to a machine significantly increases the maximum temperature. The temperature of a part made with a single laser peaked at around 162°C (324°F), but with a four-laser machine, the part reached 351°C (664°F). 

When steel that hot rapidly cools down to room temperature, it transforms into martensite, a very hard, brittle, solid solution of iron and carbon. This material also behaves differently.

In this example, a simulation revealed that using more lasers changes the quality of the part, and how it behaves in post-processes and the final application. 

Simulation allowed us to compare the maximum temperature reached in a single-laser and in a four-laser machine when printing this big part.

Example 2: Single part vs. nested build

The number of parts on a build plate significantly changes the temperature and the quality of the parts, as a thermal simulation shows.

To examine this phenomenon, we’ve compared a plate with nine parts to a plate with only one part, manufactured on the same machine from stainless steel 316L with equal settings. The difference in the maximum temperatures was remarkable, with the nested plated peaking at 248°C (478°F) and the single-part plate reaching 658°C (1216°F).

Our test showed that a single-part benchmark experiment is not representative of a production environment with nested parts. In line with this, we shouldn’t assume the parts’ quality will be the same since the temperature increase can cause further welding distortion and shrinkage and lead to different material properties.

Example simulations of a nested build job and a single part. Differences in maximum temperature are remarkable.

Example 3: Parts of different height

In additive manufacturing, it’s common practice to print tensile bars together with the part because we assume that this will ensure the two have exactly the same properties. A thermal simulation reveals it’s a flawed approach.

We conducted a simulation with eight tensile bars around one part for Ti64 material. While the tensile bar never exceeded 325°C (617°F), the part itself overheated, reaching temperatures of 600-700°C (1112-1292°F) in the upper region. Thermal simulation shows us very clearly that the much higher build rate of the upper region is the root cause. The difference in temperatures can cause a variation in the material characteristics of the part and tensile bars, even though they were built in the same process.  

This lesson applies to other production workflows. The height of parts often differs from tensile bars, which leads to a difference in maximum temperature, too. To determine the exact maximum temperature, it would be best to simulate the process.

Simulations of a part and tensile bars. While the tensile bar temperature doesn't exceed 300°C, the actual part is overheated to 600-700°C in the upper region.

These three situations are commonly encountered by engineers while working with LPBF technology. Oqton’s Amphyon helps them avoid unexpected issues with parts with an accurate build rate analysis and thermal simulation. The software also suggests an adaptation method for achieving thermally stable build jobs.