This is a collection of good, real life scientific use cases for Persistent Identifiers.
Integration with Provenance tracking
Link all components of an experiment to the corresponding Laboratory Notebook
Integration with Workflow management by identifying the workflow description and the needed data and parameters
Data publication and research validation
Long term storage, migration from disk to tape (or OpenBIS Repositories)
Substituting custom references for data fragments (e.g., database record)
Identifier for Docker images
Identify standard training data for Deep Learning (e.g., the MNIST database of handwritten digits)
Sure, there are many more examples. If you want, send us (firstname.lastname@example.org) the missing ones.