Technical Article

How 3D Printing Failures Happen in Manufacturing

January 07, 2021 by Jeff Kerns

This article will consider ways companies can use a full-digital thread and real-time monitoring that might lead to a better understanding of 3D printing part failures.

In a previous article, we looked at traditional and 3D printing failure modes. The article summarized that, "The current focus to better understand failure in metal 3D printed parts is layer orientation, process induced stresses, microvoids, crack propagation, and secondary processing." 

Understanding Failure with a Full-Digital Thread

Consistency in 3D printing is essential to moving this industry forward. 3D printed parts could show property changes from one machine or batch of powder to another. Some unsolidified powders can be reused, but mixing used and virgin powders can change materials' properties.

The idea of tracking everything through a full-digital thread is one way to learn more about mechanical properties in 3D printed parts across multiple vendors. 

One example is Materialise, a 3D printing company that produces medical implants. The company can track everything in-house, from what batch of materials end up in which parts and the location of each part. 


Materialise 3D metal printing capabilities. Image courtesy of Materialise


This way, if a problem is reported, the company can trace the root cause, go back to individual batches or powder, what machine it was built on, and more to determine what other parts might need to be monitored for the same issues. 

This lets the company alert doctors of any potential issues in other patients that received implants from the same batch of material, machine, build, etc.   

This example was for a single company, but if a database was available to track everything across multiple vendors, knowledge about 3D printing, failures, and end part characteristics could be accelerated.

IP is always a concern when organizations discuss collaboration, but using Materialise and the medical industry as an example illustrates how tracking is possible even in a highly-regulated and HIPAA compliant environment. If a patient's personal information can stay safe in a full-digital thread, surely there is a way for companies to collaborate while maintaining IP.  


Real-Time Monitoring for Metal 3D Printing 

Tracking variables such as bed and ambient temperatures, laser energy density, speed, melt pool area, hatching areas, and overlay in real-time could understand what affects mechanical behavior and failure in 3D printing. 

However, there is a lack of real-time monitoring. 

Sigma Labs is one company that offers real-time monitoring for 3D printing metals. Real-time monitoring can catch errors early, or even change process variables before a mistake happens during a print. 


Video courtesy of Sigma Labs


If an error happens, that indicates the part will not pass inspection. The software can alert a technician or directly stop the process to save materials and machine time from being wasted.

Material can be expensive, and powder can only be reused so many times. If a layer fails to meet criteria in the first dozen layers, the machine will continue printing without monitoring. 

Only after the print is finished could a technician inspect the part to find out that the last several hours of the machine's time and volume of material are now scrap. 

Additionally, automatically tracking print characteristics offers a mass of data that can be compared to the performance of the final part. Even if a part hasn't failed, it is possible to better understand over time with tracking and monitoring. 

Even with real-time monitoring and databases to accelerate understanding metal 3D printing failure, cost-effective, non-destructive testing is needed. Testing for every variable, materials, supplier, and OEM will take a large investment of time and money. Additionally, collecting field performance data or waiting for a failure to occur is also a costly and lengthy process.  


Non-Destructive Testing and Inspection for Metal 3D Printing

In a report published by NIST on 3D printing failures it reads, "The current lack of AM-specific NDI [non-destructive inspection] techniques presents a large opportunity for both research, development, and standardization. A successful NDI technique from an industrial perspective must be rapid, inexpensive, and effective at detecting critical flaw sizes. Although micro-computed x-ray tomography (µCT) is used heavily for research and validation purposes, it is too slow and expensive in its current state to be used as an everyday inspection tool. Other NDI techniques are being considered (e.g. ultrasonic)."

Much of the conversation around metal 3D printing failure revolves around process control and monitoring to reduce variation in processes and materials to predict crack propagation. 


Image showing the inspection process of 3D printed metal parts. Image courtesy of Sigma Labs


Simulation and CAD programs can predict some failure modes. More research is still needed to understand the microstructure and metallurgical properties that occur in each layer and feature of a 3D printed part. 

An article from Protolabs defines some benefits of 3D printing found in the aerospace industry and some components that have already adopted the technology.



  • "accelerating production with digital manufacturing"
  • "reducing components, which helps in lightweighting part design, cuts items in an overall assembly, and reduces costs"
  • "using a range of manufacturing methods and materials, which can accelerate development and add value to your designs"
  • "streamlining quality and compliance assurances"



  • "Heat exchangers"
  • "Manifolds"
  • "Turbo pumps"
  • "Liquid and gas flow components"
  • "Fuel nozzles"
  • "Conformal cooling channels"


Digital tools might be the only way to understand mechanical performance accurately and safety factors for each part produced to expand metal 3D printed parts into critical and fatigue applications. However, with the benefits already found in 3D printing, this research will not stop.