One-point control schemes have been the backbone of industrial distillation control for many years, although the advent of multivariate predictive controllers (e.g. Dynamic Matrix Control, DMC) has recently seen a shift towards more complex strategies. However, one-point control is still widely practised and has some inherent advantages compared with open-loop and two-point control.
One-point control is relatively easy to implement, is not subject to interactions between opposing composition control loops and provides a form of effective constraint management. Distillation columns are almost always illconditioned due to the presence of high gain variables (e.g. external flows which change the material balance) and low gain variables (e.g. internal flows which change the energy balance). If a two-point control scheme is applied, the illconditioning can restrict the attainable closed-loop performance and, in extreme cases, create instability due to excessive interaction.
The ability to implicitly incorporate constraint management into a control scheme is often more important than the composition control of the secondary product in an industrial environment. This is possible with one-point control but is often difficult with two-point control as one less degree of freedom is present in the latter case. Throughput constraints such as column flooding and reboiler and condenser duty limitations are usually nearly proportional to the internal column flows. The manipulated variable that is not used for composition control in a one-point scheme can, therefore, be fixed at a value that corresponds closely to the equipment constraint. This is particularly effective if the unused manipulated variable is the reflux rate or reboiler duty as these variables substantially determine the internal column flows.
The relationship between the unused manipulated variable and the column constraints can also be seen if it is considered that the two degrees of freedom in a distillation process determine the feed-split and fractionation. In a two-point control scheme, the feed-split controls the primary product composition and fractionation control the secondary product composition. Fractionation is a function of the internal column flows, as are the likely column constraints (i.e. flooding, reboiler duty, etc.).
The properties of one-point distillation control make it an attractive option in many cases but, although only one composition control loop must be configured, the implementation of a one-point control scheme for an ETBE column presents some difficulties. These aspects are discussed below and relate pnmanly to the selection of the controlled variable (Section 9.4.2) and the organisation of the manipulated variables (Section 9.4.3).
Was this article helpful?