1) It has to convincingly justify the CFD adopted approach.

2) It has to rely on a sophisticated computer aided engineering tool.

3) It has to prove data validity using alternative non-experimental methods such as comparative analysis.

In my opinion, a "technical paper" should adopt the best-of-the-art formulation to solve a complex flow problem giving insight into the physics. Of course, the author must provide proofs of the quality of the solution and confidence in the results.

This is one possible merit that a reviewer takes into account. Consider also that a proper journal should be considered, not all journals are suitable for technical reports.

Here are some of the features, I consider to be important to publish a CFD focused work in a journal paper:

Technical aspects:

1) Are you using a well established CFD solver/software?

2) Are the mesh and time step-size fine enough for your study?

3) Is your CFD approach well-suited for the problem you want to study? For example, if your problem involves multiphase flows or reactions, are you modeling these physics accurately?

Novelty aspect:

1) Are you able to tell something "new" to the community.

a) If you are using established CFD tools, your work should provide some insights into the physical/chemical phenomena.

b) If you are developing new CFD tools, they should improve the capability of the existing CFD solvers.