STATISTICAL STUDIES OF A REAL CONTINUOUS STIRRED TANK REACTOR (CSTR) BASED ON EXPERIMENTAL DATA

Reactor engineering is the most unique part of chemical engineering and its
technology has gain tremendous applications in industry. Three ideal contacting
patterns - batch, mixed flow and plug flow reactors are often studied and treated to
make real reactors approach ideality as closely as possible. The three reacting patterns
are easy to treat and simple to find their performance equation. Mixed flow reactor or
Continuous Stirred Tank Reactor (CSTR) is a type of reactor that is widely used
mainly in food and beverages production, chemical neutralization, and other
industries. It is preferable compared to other reactors, depending on its application for
the ease of cleaning and maintenance and requires less labor cost.
The objective of the study is to conduct statistical studies on a real CSTR. CSTR is
often treated as 'ideal reactor', in which this statement is untrue because in real life
perfect mixing in CSTR is hardly to be achieved because of dead zone creation and
channeling. This study consists of two main parts; the first part is conducting
experiments to obtain the reaction rate constant from the concentration data and tracer
analysis. The second part is to do simulation and calculation on the experimental data
obtained by using Microsoft Excel and MATLAB.
The experiment conducted is to determine the reaction rate constant of the
saponification reaction. The experiments on determining the RTD value are also
conducted. There are two methods on determining it - step and pulse input. The
experiment is conducted using the CSTR dynamics equipment and experimental data
are analyzed. From the study, it was proved that the conversion in an ideal reactor is
higher compared to the conversion in real reactors due to dead zone creation and
bypassing. The conversion for the ideal reactor is 0.496, and for the real reactor
modeled which are segregation model, maximum mixedness model and real CSTR
with dead-space and bypass model are 0.479, 0.470 and 0.480 respectively.