Upgrade of existing foul condensate stripping incorporating methanol recovery on an SA pulp mill

Stuart Bradbury, Rex Zietsman, David Rogans

Abstract
The Mondi Richards Bay Kraft mill has recently been upgraded in order to increase its overall capacity which has resulted in an increase foul condensate production. This condensate is currently stripped of volatiles (consisting mainly of methanol) in order to reduce the consumption of bleaching agent. The condensate stripping capacity of the mill has therefore had to be increased. In addition, Mondi has requested that the concentrated vent vapours be rectified to produce a methanol rich fuel for internal use. This fuel would substitute for gas in the fired lime kilns.

IMS investigated a number of possible solutions to increase the capacity including a new stand-alone stripper. On investigation of the existing equipment, it was determined that the existing tray strippers were vapour limiting, however had spare hydraulic capacity. IMS proposed to extend the height of the existing strippers and hence increase the number of stripping stages. This increase achieved the new stripping requirement from the higher liquid load without increasing the stripping steam flowrate. Since the strippers are integrated into the evaporators on the plant, this resulted in the heat balance and vapour ducting remaining virtually unchanged, and significantly reduced the total cost of the modification.

Downstream of the strippers the overhead vapour is condensed in several heat-recovery exchangers which selectively condense the water vapour, concentrating the volatiles prior to incineration. Mondi requested that this volatile stream be further rectified to produce a methanol rich liquid fuel which would have a higher calorific value than at present. The liquid fuel would be easier to store and allows surge capacity between the plants. IMS proposed a vapour fed, tray distillation column without reboiler. The condenser would control the methanol concentration in the product. The bottoms liquid is decanted to remove "red oil" (a pungent smelling form of turpentine) prior to being returned to the foul condensate strippers. The decanter is a novel design which limits the inventory of the highly flammable red oil without reducing the aqueous residence time.

BACKGROUND

In 1998, Mondi Kraft Richards Bay (MKRB) embarked on a mill expansion programme culminating in an extended shutdown to install and commission new equipment in August, 1999.

As part of the expansion, the foul condensate flow rate through the mill increased by 10 l/s. This necessitated an increase in foul condensate stripping capacity to treat the additional flowrate.

In addition, MKRB specified that the concentrated vent vapours to be rectified to produce a methanol rich fuel for internal use. This liquid fuel would partially replace the expensive gas currently combusted in the lime kilns.

IMS investigated a number of possible solutions to increase the stripping capacity. These alternatives considered capital costs, operating costs, layout considerations, and the practicality of implementing the changes during the short planned shutdown.

Prior to discussing alternatives, it is worth reviewing basic distillation theory as an aid to understanding the implications of the different parameters on stripper column design. This will assist in explanation of the decisions taken.

1. A BRIEF REVIEW OF STRIPPING THEORY APPLIED TO TRAY COLUMNS

In tray stripping columns, liquor (in this case water) containing a volatile component is fed to the top of the stripper and allowed to cascade trays by tray down the column until it reaches the bottom. Stripping vapour (steam) is introduced below the bottom tray rising upwards through the trays. The vapour and the liquid phases are intimately mixed on each tray resulting in an approach to thermal and mass transfer equilibrium during which the volatile component is preferentially transferred from the liquid to the vapour phase.

The number of trays required in a particular system is determined by the relative volatility of the volatile component compared to the transporting liquor. This relative volatility determines the extent of the transfer of the volatile component at each equilibrium or "theoretical" stage. In general, the higher the relative volatility of the volatile component, the fewer theoretical stages or trays will be required. It is thus possible to calculate the number of theoretical trays required to achieve the desired stripping, and achieve the desired bottoms outlet concentration. There are several methods for calculating the theoretical number of trays which are well described in most mass transfer textbooks.

For description purposes, the McCabe Thiele graphical method has been selected to illustrate this process. In the McCabe Thiele method, an "operating line" is determined based on a mass balance on the column. An end point is chosen for the "bottoms" (i.e. the maximum acceptable concentration of volatile component in the stripped liquor). The slope of the operating line is then calculated, based on the liquid to vapour (L/V) ratio. Graphically, the operating line is drawn through the outlet concentration with a slope of L/V. The number of theoretical trays required is then obtained by "stepping off" between the operating line and the vapour liquid equilibrium concentration curve.

It may be observed that the greater the slope of the operating line (L'/V>L/V), the higher the number of trays required to achieve the removal of the volatile component to the desired outlet concentration. The actual number of trays for a desired separation is dependant on the number of theoretical trays and the individual tray efficiency.

In summary:

The number of theoretical trays required for a particular stripping system is determined by the desired outlet concentration of the volatile component in the stripped liquor and the L/V ratio.

2. A BRIEF REVIEW OF COLUMN DESIGN

It must be borne in mind that the hydraulic performance of a column is independent of the stripping design and stripping efficiency of that column.

Flooding

The diameter of a tray column is determined by calculating two important criteria:

  • Liquid flood, which is the ability of a column to cope with the liquid flowing down through it, and
  • Jet flood which is the limit to vapour flowrate through the column at acceptable pressure drop without entrainment of liquor droplets from any tray to the tray above.

If either flooding condition is exceeded, the column will operate under conditions of higher pressure drop whilst excessive entrainment of droplets from tray to tray will nullify the separations achieved. Hence no appreciable stripping will take place. These criteria are interdependent with guidelines for minimum and maximum rates. It is generally accepted that 80% to 85% of both jet flood and liquid flood are the maximum rates that should be used in normal design.

Foaming

Another important consideration in hydraulic design is the amount of foaming anticipated. Foaming has serious consequences on jet flood since it may result in the vapour entraining liquor to the tray above. This again nullifies the separation achieved on the tray. The overall system will thus rapidly lose efficiency. The correction factor for "mildly foaming" was used in this case as minor foaming conditions are occasionally experienced in foul condensate strippers.

Tray Efficiency

To determine the number of actual trays from the number of theoretical stages, tray efficiencies must be calculated. Tray efficiencies are affected by many parameters and use of an "average" tray efficiency to calculate the number of actual trays is not good practice. For example, in the methanol column, the average tray efficiency is 65% while in the strippers the average tray efficiency is approximately 40%.

Tray suppliers understand the operation of their products, and are therefore best suited to guaranteeing their performance and specifying the individual tray efficiency.

3. EVALUATION OF THE EXISTING PLANT

Before embarking on designing a new plant, it was decided to evaluate the existing plant to ascertain if the additional capacity could be achieved from the two existing strippers.

It was assumed that the existing stripping columns were hydraulically able to cope with the additional 10 l/s of foul condensate. If the foul condensate flow increases while keeping all the other parameters fixed, the L/V ratio will also increase. From the theory above, this increased the slope of the operating line approaches the equilibrium curve, necessitating more trays for the same bottoms concentration. As the number of trays in the existing strippers was fixed , the current stripping efficiency would decline. The methanol in the bottoms consumes bleaching agent and hence this was unacceptable to Mondi.

In order to reduce the bottoms concentration of the existing strippers, the vapour flowrate to the strippers would have to be increased by L'/L (in order to return the slope of the operating line to its original slope). If this could be achieved, then no additional stripping equipment would be required.

On this basis, the hydraulic calculations were reviewed. The calculations indicated that both stripper columns could barely accommodate the additional foul condensate without flooding while maintaining the same stripping vapour flowrates. Jet flood on the first column rose to 85%, the maximum allowable, while on the second column it rose to 65%. This indicated that the second column had some capacity to increase the vapour flowrate. Unfortunately, any additional vapour added to this column would be condensed downstream of the strippers and returned as reflux. This additional liquid would result in the towers operating in the liquid flood range. Consequently, it was not possible to use the existing strippers unchanged to accommodate the additional foul condensate flowrate while still maintaining bottoms concentration. It was thus concluded that more trays were required in the stripper columns.

4. INCREASING THE NUMBER OF TRAYS

There were various options considered to increase the number of trays in the stripper columns. These were:

  • To physically cut the tops (or bottoms) off the existing columns, lift in a new section with additional trays and weld the system together,
  • To install a new independent stripper on the ground that would act like a new "top" section to both existing strippers,
  • To install a new independent stripper on the ground that would act like a new "bottom" section to both existing strippers, and
  • To install a new stand alone stripper to cope with the additional foul condensate flowrate, leaving the existing system unchanged.

There were advantages and disadvantages to each option. A cost exercise was undertaken and the most cost-effective solution found was to cut the top dish off the existing strippers and to add an additional 13 trays to each column.

The most significant benefit of this option was that the vapour flow through the columns was unchanged. Consequently the vapour flow to and from the evaporator did not change and thus their operation would be unaffected. The volume of stripping steam would also be unaffected by the increase in duty suggesting that this option was favourable when considering operating costs. Furthermore, the existing vapour ducts could be used without modification leading to large cost savings since these ducts are all manufactured in stainless steel.

Feed Preheat

Another consideration was the foul condensate feed temperature. If the feed liquid was below boiling point, some of the vapour flowing through the feed tray would be condensed. This would increase the liquor flowrate down the column below the feed tray and reduce the amount of vapour to the tray above. Consequently, steam driven feed preheaters were installed to ensure that the feed was at boiling point on entering the columns.

5. SOLUTION TO THE NCG PROBLEM

Downstream of the strippers, the stripper overhead vapours are condensed in several heat exchangers mostly associated with the evaporators. A water cooled trim condenser maintains the final vent gas temperature. The condensers concentrate the vent gases. In the existing plant these gases are currently ducted to an incinerator where they are burnt with gas to an odour free state.

Mondi utilises gas fired lime kilns in the overall reagent cycle. These kilns operate with Sasol gas as the fuel. A major incentive with the optimisation project was to reduce the cost of fuel for the kilns. Mondi decided to install a methanol plant to produce a methanol rich fuel. This would reduce the amount of gas fired in the incinerator as well as on the kilns.

It was decided to install a vapour fed methanol rectification column to concentrate the NCG's from about 30% v/v methanol to greater than 80% w/w methanol fuel. This fuel would comprise mostly of methanol as well as other low boiling point organics present in the NCG's.

By selecting a vapour fed column, the need for a reboiler was eliminated. The bottoms from the column are sent back to the strippers as reflux. In the strippers the methanol is stripped out and returns to the methanol plant once more via the vent gas system. The capital and operating costs were significantly reduced by not having to install a reboiler on the methanol column .

Furthermore, the column was fitted with an overheads partial condenser controlled by the column temperature profile. The benefit of this is that the column is virtually insensitive to the incoming vapour concentration as the reflux is adjusted to maintain methanol concentration at all times. This was particularly useful as the fuel to the lime kilns would be kept at a consistent concentration.

In the concentrated vent gas there is a small quantity of pungent smelling turpentine oils collectively known as "red oil". This oil has to be removed from the system to prevent build-up in the circuit. To remove the "red oil" IMS installed a novel decanter. During the "hazop" study on the plant, Mondi requested that the inventory of "red oil" be minimised as this oil is highly flammable. Consequently, the traditional horizontal cylindrical decanter was rotated into a vertical plane. The feed and discharge nozzles were positioned so that three hours of aqueous residence time at the bottom of the tank was provided with less than 200mm of oil floating on top of the aqueous phase.