These results illustrate the improvement in dynamic and steady-state performance that is achievable with conventional PI control structures through the use of ratio and cascade control schemes. Of course, online composition measurements are required for dualcomposition control.
This process provides an interesting and important example of the conflicts between individual unit performance and plantwide performance. Considering only the column in isolation, control structure CS1 gives the best results in terms of product purities for disturbances in both feed flowrate and feed composition. However, this control structure produces large and rapid changes in the vapor distillate flowrate, which could seriously degrade the performance of downstream units. Control structure CS3 provides more gradual changes in the distillate flowrate, but it does not hold product purities as close to their specifications as control structure CS1 does in the face of feed composition disturbances.
Modifying the CS3 structure by the addition of dual-composition control provides effective product quality control with essentially the same low variability in distillate flowrate.
A final comment might be useful in situations where the coolant is a boiling liquid (e.g., a refrigerant such as propane). The heat transfer rate in this situation depends on the pressure and composition of the coolant and on the coolant liquid level (since this affects the heat transfer area). Normally the level is maintained to keep all the tubes covered. The coolant pressure is adjusted to control the heat removal rate, which can be conveniently determined by simply measuring the flowrate of the coolant vapor leaving the condenser. Thus this more complex system with cooling-side dynamics can be handled in a straightforward way, and the performance of the alternative control structures should be similar to that found in this study in which the heat load is assumed to be manipulated directly.
Was this article helpful?