SIMULATION OF FLAT FALLING FILM EVAPORATOR NETWORK

Abstract

The present investigation is aimed at modelling the flat falling film evaporator (FFFE) evaporator network employed for the concentration of black liquor in aPaper Mill, for the recovery ofchemicals. As a base case, the data ofa nearby Paper Mill employing a seven effect FFFE evaporator network, of which the first two effects are finisher effects employing live stream, with backward feed sequence is considered. It is a well-known fact that evaporation is a energy intensive process and the evaporator house of a Paper Mill consumes about 24 to 30 percentage (Rao and Kumar (1985)) of the total energy consumed in a typical Indian Paper Mill. The efficient use of a evaporator network depends on many factors such as proper selection of operating parameters, which include live steam temperature, last effect temperature, feed sequencing, feed -, product -, condensate - flashing, feed splitting, vapour bleeding and use of re-heaters in proper places. Most of the modelling efforts in past has been concentrated towards modelling a evaporator network, with the help of usual mass and energy balance equations, where heat transfer coefficient is an input parameter. Many authors also developed empirical or mathematical model for the prediction of overall heat transfer coefficient based on different operating parameters of the effect. When the above two models are combined, it gives an effective and efficient model to study the effect of operating parameters of the evaporator network. Most of the past studies have been concentrated on equation based models and are around long tube vertical (LTV) evaporators. The equation based models are comparatively rigid in their framework, in dealing with complicated operating environment. In fact, it is difficult to provide flexibility in these models and its translation to computer codes through data set to include different flow sequencing, flashing, splitting, and bleeding provisions. In the present work, a mathematical model has been developed to incorporate the above operating modes with an aim to figure out an efficient strategy for the operation of FFFE. The mathematical model has been translated to a FORTRAN code, and is run using a Pentium-IV machine. The results of the mathematical model compares well with the operating parameters of the base case. This validated model then, has been used to study the different combination of operating parameter, with an aim to improve the steam economy of the base case. The different feed sequences, as employed in Scandinavian countries along with backward and various mixed feed sequences, has been investigated. The base case doesn't have re-heaters in the network. The effect of re-heaters has also been investigated, to improve the steam economy of the network. It should be noted that re-heaters draw vapour from the network itself, and not from any external source. The present work is based on the work of Stewart and Beveridge (1977), and Ayangbile, Okeke and Beveridge (1984). It provides a flexible framework of modelling, which can accommodate different operating alternatives. The above model has been improved and modified to accommodate different operating alternatives such as - steam splitting, provision for condensate -, feed - and product - flash, feed/liquid splitting, vapour bleeding for re-heaters etc. Many empirical correlations has been developed such as correlation for BPR, overall heat transfer coefficient of flat falling 11 film evaporators, and few other physico-thermal property corrections. Using these correlations and mass and energy balance around an evaporator a cubic equation has been developed. It's one ofthe real roots provide the exit liquor flow rate. The above developed model has been able to predict the concentration profile, and temperature profile within a error bandof+2%. Once the model has been validated with the plant data, it is used to find out the effect of different operating alternatives towards steam economy - a parameter used a yardstick to measure effectiveness of the FFFE network. The effect of different operating parameters on steam economy has also been included in the present work. From the results obtained, it has been clear that backward feed provides the best steam economy out of the flow sequences investigated. A comparison between the investigated sequence, and that of Scandinavian sequence shows that the later is inferior as regards to steam economy. The base case operates in backward feed sequence mode with steam economy of 5.0. However, when re-heaters are employed with the base case FFFE network, there is a noticeable improvement in steam economy, which rises to 5.66 registering a growth of 13.1%. This amounts to a gross saving of steam equal to 5660 t/y. It appears that the feed temperature of the base case is low (at about 65°C) which can be increased to 85°C to improve the steam economy. Further it has been observed that if the FFFE network with re-heater is operated by keeping steam temperature, last effect temperature, feed temperature, feed flow rate and feed concentration equal to 120°C, 52°C, 85°C, 56200 kg/h and 0.118 respectively, the steam economy will increase to 6.12. This will amount to a saving in steam consumption equal to 11650 t/y.