Experiment-Based Decision Making in Complex Systems

Managing complex systems is a challenging task as decisions are much more difficult to make and consequences (in case of wrong decision) much harder to bear. The process industry falls certainly in the category of complex, safety-critical systems. Disasters like the “Deepwater Horizon” (2010, 11 fatalities), the BP Texas City (2005, 15 fatalities), and the AZF Toulouse (2001, 29 fatalities) witness that. The paper explains the advantage of grounding decisions on the outcomes of realistic simulations performed in a Plant Simulator (PS) and also how, in this way, operators’ competency can be better managed and human performance improved. The results of the first experiment performed in a PS (the first known in the process industry to date) are used to clarify the concepts explained. They help clarifying how conveniently the data collected in a PS can be used for making daily decisions on how to better run the plant (safely and more efficiently).