The classical Design of Experiments (DoE) methodology has limited itself for many decades, designing experiments whose conditions do not change with time. It is very likely that the optimal operating conditions, say of a batch process, might be changing with time and the classical DoE experiments will help us discover such optimum. We will only arrive at a suboptimal operation with an operation that does not change with time. To discover the time-varying optimal operation, we need to design time varying expariments
Dynamic Experiments Provide Important Information
Operating Conditions Change with Time
Experimental Domain
Simple Examples of DoDE Experiments
The two sets of experiments with linear dependence on time (Left) and quadratic dependence on time (Right)
Approximation of profiles for ease of implementation
With Linear Segments or Step-wise Changes
For operational ease, the time-dependent operational profiles can be approximated by a series of linear segments (left) or steps (right). The selected approximation should be the same in all experiments.
DoDE experiments with varying duration
DoDE experiments can be efficiently designed when the duration of the experiments is also a factor.
The combination of time-varying factors, called dynamic factors, and time-invariant factors, called traditional factors, is straightforward.
