An industrial case study is when you take your solutions or your processes, apply it to some situation in industry, and then report on the results. You might use the Case Study Method research method if what you propose can be appropriately investigated using a qualitative case study. This is appropriate if you're in management sciences, human-computer interaction, and software engineering (which is what the Damian paper referenced above is about). Note that industrial case studies of this type are VERY difficult to do and take an extremely long time to write up and report on, so I actually doubt the conference is looking for this kind of case study.
Alternatively, in the field of Knowledge Discovery of Databases, an industrial case study may simply be applying your algorithm to a database that is used in industry to discover something novel or meaningful. This is more common for analytical fields in general and simply requires you to acquire a database or some other data source that was generated through industrial activity.
The reason for "industrial case studies" is because a lot of people create methods and techniques and test it only with "fake laboratory data" that they happen to create or acquire. This fake laboratory data doesn't reflect very accurately the real world and people want to see that techniques in fact work on, affect and influence the real world.