Objective of the Experiment
Authored by: James N. Cawse, Ph.D.;Cawse and Effect LLC
At this stage of the program, the focus should be on diagnosis, not on solutions. A good problem statement avoids looking to the solution prematurely, and helps make the right problem is solved.
- What is already known about the system?
- Have the tools of process mapping, root cause analysis, and Pareto Analysis been used to make sure the right things are being studied?
- Will some business action result from this experiment? (If not, why are you wasting your time??)
- Is this Exploration or Exploitation? Is this just at the beginning of the study, trying to find out what factors (control variables) have a strong influence on the system? Or, is there basic knowledge of the system and the objective is optimization?
- Is it essential to get the properties to their maximum levels or is consistency more important?
- Unbiased: the team must encourage participation by knowledgeable and interested people with diverse perspectives.
- Specific and measurable: the objectives should be detailed, preferably quantitative, and stated so that it is clear whether they have been met.
- Of practical consequence: something will be done differently as a result of this experiment.
Relevant Background Knowledge:
What are the known theoretical relationships, practical knowledge, and results of previous experiments? The purpose of this information is:
- To establish a context for the experiment and a clear understanding of what new knowledge can be gained.
- To motivate discussions about the relevant knowledge.
- To uncover experimental regions of particular interest and regions that should be avoided.
Every effort must be made to get maximum participation from all involved parties - the experimental team, management, and customers.
An HT experiment is very different from a conventional DOE project, where the total number of runs may be only one or two dozen. Any HT experiment worth starting will have have hundreds to thousands. It is critical to get an initial estimation of the size of the experimental space that is being studied, and compare it to a reasonable budget of runs that will be applied to it.
Experimental space is an appealing intuitive concept, but it is difficult to pin down. It is sometimes defined as the product of the number of levels of each of the variables. That can easily balloon to impressively large numbers, but such calculations typically founder when their assumptions are tested. For example, how many level are assumed for a quantitative (real) factor? What level of interactions need to be searched in the qualitative variables? What constraints are present in a mixture space?
Nevertheless, a rough estimate of the total number of reasonably distinct experimental points is possible. From that it is possible to estimate if the runs budget will lead to very sparse or relatively thorough coverage. This has important implications to the final design selection.
|Click [+] for other articles on||The Market Place for Lab Automation & Screening||Automation Software|