Optimisation Results

6.3.3.1 Steady-State Solutions

The results of the optimisation are shown in Table 6.10. Two additional optimisation runs were performed to demonstrate the incentive for considering both reaction and separation together. The maximum profitability is achieved when neither the ETBE purity nor the isobutene conversion is at its maximum. Underlined results indicate active constraints.

Tabic 6 10 - Optimised Process Parameters and Results.

Maximum Value

Maximum

Maximum

Added

ETBE Purity

Conversion

Set-Points

Overhead Pressure (kPa)

650

725

750

Stage 7 Temperature (°C)

104

113

105

Reflux Rate (mVhr)

17.5

21

22

Key Results

ETBE Purity (wt%)

95.1

95.7

95.4

Isobutene Conversion (mol%)

94.6

95.4

95.5

Profitability ($/m3)

1744

1730

1733

Constraints

Reboiler Duty (MW)

5.10

5.64

5.73

Condenser Duty (MW)

3.90

4.35

4.41

Maximum Flooding Factor (%)

74

80

80

Ethanol in Distillate (wt%)

0.18

0.38

0.25

C< in Bottoms (wt%)

0.45

0.36

0.63

The profitability of the unit is limited by the maximum condenser duty. A decrease in the column pressure would increase the net value added by the process if the overhead vapour could be condensed at the lower temperature. This indicates that there is an opportunity to increase profitability by debottlenecking the condenser until one or more additional constraints become active. Interestingly, a higher ETBE purity and isobutene conversion are achievable at pressures above the optimum. The increased pressure alleviates the condenser limit but introduces a flooding limit.

It is pertinent to examine the unconstrained extrema (i.e. peak values without constraints) in this system. The maximum unconstrained profitability of $1750/hr is achieved at a pressure of 500 kPag. At lower pressures, the departure from chemical equilibrium increases due to the slower reaction rate and the maximum conversion (and consequently the maximum purity also) decrease. The maximum unconstrained ETBE purity is 97.3% and occurs at very high internal vapour and liquid rates. The profitability at this point is only $1470/hr due to the excessive use of utilities. There is no maximum unconstrained isobutene conversion as increasing the reflux rate (and adjusting the other parameters appropriately) always increases the conversion.

6.3.3,2 Supervisory Control System

The results of this process optimisation can be used within a supervisoiy control system. This is shown in Figure 6.5 which considers the dynamic response to set-point changes implemented by a supervisory controller following a process disturbance. In this case, the disturbance was assumed to be a decrease in the feed rate from 120 kmol/hr to 100 kmol/hr. Initially, the column is operating with the optimal set-points and, therefore, maximum profitability. The column is allowed to stabilise at the new feed rate before time = 0 in Figure 6.5. Thus, the set-points are now sub-optimal and the column is operating below peak profitability.

The supervisory controller calculates new set-points by optimising the column operation at the new conditions. These set-point changes are implemented at time = 0 in Figure 6.5 and have an immediate destabilising effect on the column. Initially, this reduces the profitability but after the process transient is completed, the profitability has been increased. The net economic benefit of employing the supervisory control system is equivalent to the area between the solid line (supervisory control system implemented) and the dashed line (set-points not updated after disturbance). After less than two hours, the initial lost production has been recovered and, thereafter, the supervisory controller makes a net profit.

Figure 6.5 - Process Response to a Feed Rate Disturbance: Benefits of Supervisory Control

The supervisory controller can also be used to optimise the process performance following feed composition changes. However, steady-state simulations show that there is little incentive to manipulate the process set-points in response to feed composition changes: the net benefit is less than $ 1/hr for compositions in the practical interval around the base case. This is convenient as it is far more difficult to implement a supervisory control system which must reflect compositional changes as well as rate changes. Process analysers are required to measure compositions and these introduce an extra cost and process dead time. This dead time reduces the process controllability and results in a more sluggish response to disturbances. Consequently, the viability of the compositional aspect of a supervisory controller is questionable.

Was this article helpful?

0 0

Post a comment