Basic principles of HDRI
A high dynamic range image (an HDR) is created from three or more impressions of the exact same scene. That’s three physical camera clicks (each image taken with different exposures), and three image files combined into one using HDR software. It’s possible to duplicate one photograph and edit each to produce three source files to input into your HDR software; that’s not a true HDR.
The technique may be used to artistic effect, or for technical reasons. The best HDR images are constructed according to a rigorous technical process, and the same tenets of post production should define the aesthetics when it comes to combining the photographs.
If HDRI is used to fulfill a technical objective (for example, showing the detail a viewer sees through a window in a room which would otherwise be washed-out on camera), the three images used must appear as close to carbon copies of one another as possible. No movement of objects in frame must occur, however small. For this reason HDRI is best suited to static scenes, and it is unsuitable for portrait photography.
It must be made using a tripod to ensure absolutely no movement of frame. Even a moving tree in the distance can ruin an HDR image.
If HDRI is used for aesthetic effect, movement can add to the interest, especially where the source images are long exposure, or where it’s desirable to trace the movement of an object (e.g. light trails, moving crowds, etc.)
HDRs are usually produced from three RAW images, which were made at exactly two stops apart from one another. (-2EV, 0EV, +2EV).
Source : http://www.matsmithphotography.com/photolife-blog/hdr-principles
These principles can be applied in hundreds of different ways.