Total Reflux Tests

A senes of tests was completed to determine the maximum reflux rate for various values of the reboiler voltage. The reboiler sump was filled with approximately 10.0 L of liquid (mostly ethanol) before the reboiler elements were turned on. The feed pump was stopped and both the bottoms control valve and the distillate control valve were fully closed. The chilled water flow to the condenser was maximised (i.e. approximately 50 L/min) and the reflux control valve was opened fully.

The column pressure was permitted to increase until the rate of condensation in the condenser exactly balanced the rate of vaporisation from the reboiler (i.e. the process reached steady state). At this time, the reflux rate was recorded via the flow meter and the condenser duty estimated from the chilled water inlet and outlet temperatures. This data provides a secondary measurement of the reflux rate to verify the flow meter reading which was suspect due to an intermittent instrumentation error. The reflux rate was also estimated via simulation (Pro/II) from the estimated reboiler duty which was calculated from the voltage by equation (12.3). Three different reboiler voltages, 100 V , 150 V and 180 V, were tested in this way with the results shown in Table 12.12.

Table 12.12 - Results of Total Reflux Tests

Reboiler

Condenser

Reflux Rate

Voltage

Est. Duty

Temps

Est. Duty

Pressure

Meter

Calc.

Sim.

(V)

(kW)

(°C)

(kW)

(kPa)

(L/min)

(L/min)

(L/min)

100

3.2

9.0-10.5

5.2

390

n/a

0.51

0.30

150

6.0

9.9-12.4

8.7

460

0.56

0.86

0.58

180

7.9

7.9-9.5

5.6

840

0.61

0.61

0.84

The three different measurements of the reflux rate correspond reasonably but there are significant uncertainties associated with each reading: the meter reading should be accurate to within 1% but the display periodically resets to zero when there is a known flow of liquid so that the reading is biased towards zero (a filter was used to stabilise the reading); the calculated reflux rate is sensitive to slight changes in the chilled water temperatures and the uncertainty in the chilled water flow rate was estimated to be 30%; the reflux rate determined via simulation is subject to errors in the empirical reboiler duty correlation and heat losses from the column and was estimated to be correct to within 25% only. Additional instrumentation (e.g. a more reliable flow meter display unit, accurate measurement of the chilled water flow and a power indicator for the reboiler) would permit a more accurate determination of the reflux rate.

The latent heat of C4 hydrocarbon is only 30-35% of the latent heat of ethanol. Therefore, a reflux rate of greater than 2.0 L/min is predicted for the expected operating conditions with a reboiler voltage of 180 V. This is acceptable since it permits a reflux ratio of approximately five to be investigated at the proposed feed rate (0.5 L/min), and higher reflux rates and reflux ratios could be investigated with an increased reboiler duty. However, the reflux rates and condenser duties calculated from these tests were significantly lower than anticipated from the condenser performance with water (Section 12.1.6). The relatively high pressures that were observed in the total reflux tests are also apparently inconsistent with the predicted condenser duty. Further tests (e.g. GC analysis of the feed to confirm the concentration of light components) are required to fully determine the source of the discrepancy but the apparent poor thermal performance is mostly due to differences in the physical properties between the test conditions and the expected operating conditions, as described below.

Ethanol has a much lower Prandtl number and a lower thermal conductivity than C4 hydrocarbon and these differences result in much lower heat transfer rates. The effect is compounded by the low condensation rate of ethanol which reduces the Reynolds' number and, therefore, produces a lower Nusselt number inside the inner tube. This combination of effects reduces the inner tube heat transfer coefficient by approximately 70% compared with the design conditions. Since the heat transfer rate is controlled by the inner tube heat transfer coefficient and fouling, the influence of the physical properties reduces the overall heat transfer coefficient by 90-95%. Additionally, the flow regime was transitional for these tests and this affects the accuracy of the heat transfer correlation as indicated above (Figure 12.5). A greatly increased condenser duty and, therefore, lower operating pressures are anticipated for the expected operating conditions (i.e. high concentration of C4, etc.) due to these factors.

Chapter Thirteen Conclusions and Recommendations

Was this article helpful?

0 0

Post a comment