Quality Assurance for Injection Molding

Reliable Production: Quality Assurance for Injection Molding

Resource-friendly production and minimization of rejects save companies time and money. High quality enhances reliability and ensures components have a long service life. Fraunhofer IFAM has promising new results for achieving these aims for metal injection molding.

On the way to zero-fault production


Metal injection molding has become an established technology for large series production. It is therefore vital for industry to minimize rejects at every processing stage (feedstock, injection molding, binder removal / sintering). In order to achieve the goal of zero-fault production:

  • Relationships between feedstock homogeneity, rheology, and the ability to effectively simulate the mold filling process are being studied, so allowing production-ready molds to be manufactured more rapidly
  • Neural networks are being used to separate defective components directly after the injection molding
  • Direct interactions in the process between the sintering atmosphere, binder components, and metallic powders are being analyzed in order effectively adapt the material properties and optimize the sintering cycles

Use of neural networks in injection molding


The dimensional stability and quality of the sintered component highly depend on the selected parameters for each preceding processing step. Deviations in the process parameters lead to dimensional variation and in the worst case to component defects such as cracks, deformation, and cavities which are only detected after sintering. The aim is to acquire information during the injection molding process about the quality of the components and the effect of deviating process parameters.

One approach for this is the issuing of a quality statement solution synchronously with the production cycle, with full documentation of all relevant information. This is being developed with the help of neural process control systems. These even allow control of very complex relationships in a self-learning and constantly improving system.

An instrumented injection mold allows the following properties to be tested and predicted:

  • Tendency to separate at the flow front
  • Weld line sensitivity
  • Feedstock flowability (maximum flow distance 510 mm)
  • Pressure and temperature drop across the flow distance
  • Charge homogeneity
  • Reproducibility of the injection parameters