Jari Kapanen and Lasse Kuusisto

Company and address

Metso Paper Automation Inc., Tampere, Finland


lasse.e.kuusisto@metso.com and jari.kapanen@metso.com


brown stock washing, process control, washing models, automation, optimization controls


A new method of achieving stable control of brown stock washing lines has been implemented in four kraft pulp mills in Finland. This method is based upon the development of a new model of the washing process called Washing Potential in Diffusion. This model has been adapted for the control of several commercial brown stock washers.

Once the stability of the individual washers has been achieved, higher level optimization controls then produce stability in the entire washing line during normal production and during production rate changes and during wood species changes. The optimization control programs achieve more stable and increased washing efficiency resulting in lower washing losses, reduced loading on evaporation plants and higher pulp outlet consistencies. This technical paper describes the development and implementation of a brown stock washing optimization system. Results from kraft mill production lines are presented.


Until recently, the brown stock washing area has not been highly automated and its performance not been optimized in any significant way. In most pulp mills the brown stock washers are equipped with basic instrumentation and the individual control loops are supervised by a DCS system. The coordination of these individual loops is done manually by the operators. In any manual operation the results vary with the individual so the stability of the operation is not optimized in many cases.

But many mills are now realizing that poor performance of the brown stock washers can have a significant impact on the stability of operation and the production costs in other related processes. For instance, poor washing of the pulp can result in higher bleaching chemical consumption. Organic materials that are not washed out of the pulp are a lost source of energy for the recovery boiler. On the other hand, if the filtrate liquor is too dilute, evaporation costs will be higher. In some cases this may result in a production bottleneck for the whole pulp mill. Also, if outlet consistency is not regulated and becomes too high the washer operation may be halted.

The operation of the brown stock washing operation is therefore a balancing act; with the operators trying to achieve the best removal of soluble impurities, the highest possible outlet consistency and the highest solids content in the dilute liquor sent to the evaporation plant. But operating practices vary from operator to operator and, in some mills, the wood species and production rates change on a regular basis, thereby destabilizing the washing process. Because the washing line is a sequential, countercurrent operation, the stability of the whole line can be upset. It may take only a few minutes to destabilize, but the recovery may take several hours.


Some pulp mills are now realizing that poorly managed brown stock washing increases the cost of pulp mill operations. It is therefore any area of opportunity for cost savings or for removing production bottlenecks. The potential benefits of brown stock washing line optimization are shown in Figure 1.

Figure 1

Figure 1: Benefits of washing line optimization

The benefits are derived from better stability of the washing operation then optimized performance. With a consistent, coordinated control looking after the operation of the whole line, differences in operating procedures are eliminated and upset conditions are handled quickly and in a coordinated way. The whole washing line is balanced so that the washing load is effectively distributed between all washing units.

A more consistent and thorough removal of soluble impurities results in reduced chemical consumption in subsequent pulp bleaching stages. More consistent dilute liquor with a higher average solids content will save evaporation steam costs. A higher outlet consistency of the washing line will decrease effluent load and allow a better division of alkaline and acid washing water in the bleach plant. More efficient washing also has a positive impact on pulp quality. For instance, if impurities are effectively removed pulp fiber properties will not be degraded in the oxygen delignification operation.


Obviously, for any high-level, hierarchical control to take place the process must be appropriately instrumented and controlled by a DCS system and all loops must be well tuned. The optimization process has three levels as shown in Figure 2. This figure shows the hierarchy of controls for a Drum Displacement (DD) washer.

Figure 2

Figure 2: Hierarchy of optimization controls for a drum displacement washer

First, it is important to note the brown stock washing line cannot be effectively optimized unless the individual washing stages are operating in a stable fashion. The first step of optimization is therefore at the washer level. Washing models, specific to a type of washer, are employed to stabilize their operation and to achieve the highest possible washing efficiency in every washer in the line. Secondly, the optimization of the entire washing line includes control of the washing factor and other material balances throughout the washing line.

The washing efficiency of each washer is optimized using a special control method called Washing Potential in Diffusion, described later. Then, the washing factor in the whole washing line is optimized by dividing the washing load between the individual washers. Finally, the usage of washing water is optimized by dynamic adaptation so that the capacity of the evaporation plant is used effectively, without overloading it.

During fiber species changes - from softwood to hardwood – washing liquor is stored in filtrate tanks at the end of the softwood run so that enough wash water is available for the increased production rate during the hardwood run. When changing from hardwood to softwood with a lower production rate, a reverse strategy is used.

Now, we shall describe the mechanisms of individual washer optimization controls.


There are three basic mechanisms involved in the washing of pulp. See Figure 3. These are dilution/extraction, displacement and diffusion.

In dilution and extraction washing, washing water is mixed with the pulp and the solute -containing liquor is then removed by drainage through a wire or by pressing. In displacement washing the washing water ideally displaces an equal volume of liquor in the pulp. In practice, some mixing of the wash water and displaced liquor takes place. Different washers use a combination of these mechanisms.

Figure 3

Figure 3: Mechanisms involved in pulp washing

The efficiency of the dilution and extraction mechanism depends mainly on the consistencies to which pulp slurry is initially diluted and finally thickened. The pulp slurry should be diluted as much as possible and the pulp consistency after thickening should be as high as possible. To enable longer retention times and thus a higher degree of diffusion, the dilution and extraction process should be slowed down to remove solute sorbed within in pulp fibers.

Retention time the in the displacement process is quite long, which increases diffusion. Diffusion is the main principle used to remove solute that has been sorbed by cellulose fibers in pulp. The displacement process is dynamically slow and thus retention time for diffusion becomes an important element in efficient washing.

In brown stock there still remains quite a lot of sorbed solute, both organic and inorganic, so the diffusion effect become vital in washing, regardless of the mechanism. As a starting point for the optimum control of a washing process, regardless of its design, a washing model has been developed to explain how soluble impurities are removed from fibers by the diffusion process. These soluble materials are then carried away in the extracted liquor. This mass transfer model is called Washing Potential in Diffusion. In simplified form, it is expressed as a functional relationship.

Q= f (C, t, T)  

Q = Washing Potential in Diffusion

C = Consistency

t = Time

T = Temperature

Consistency in this equation refers to all consistencies in the washing line – pulp consistency, dry solids consistencies and solute consistencies. More effective washing is achieved with a higher difference between pulp inlet and outlet consistencies. Similarly, more effective washing is achieved if there is a higher difference in solids content between the wash water and the extraction liquor.

The washing potential defines the capability of any washer to remove solute from the fiber. This depends on the level of solute within the fiber compared to that in the wash liquor. Time refers to the residence time during which the fiber and the wash liquor are in contact. Since temperature affects the mass transfer rate it should be measured but it is not controlled in practice.

This basic washing mechanism is used for controlling the washing process in several different washing processes, all of which depend on diffusion. Applications of this model include:

  • Drum Displacement (DD) washers built by Andritz
  • TwinRoll Presses built by Metso Paper
  • Atmospheric two-phase diffuser washers built by Andritz or Kvaerner
  • Drop leg washers
  • Pressure diffuser washers built by Kvaerner
  • Wash presses built by Kvaerner


The Washing Potential in Diffusion model, so-called soft sensors and fuzzy logic controls have been applied to the Drum Displacement (DD) washer manufactured by Andritz. Other washer-specific models have been developed. They are not described in this paper.

The DD washer is multi-stage washer that enables high washing efficiency. But the complexity of the washing sequences places extra demands on the controls. The basic process instrumentation required for this type of washer control is shown in Figure 4.

Figure 4

Figure 4: Instrumentation required for DD washer control

The measurements required for DD washer controls are shown in Table 1:




Washer inlet, outlet, washing and extraction liquors

Pulp Consistency

Before and after washer

Tank levels

Mass tanks (blow tank, stand pipes) and filtrate tanks


Extraction flow or in filtrate tank


Feed pressure

Dry solids content

Black liquor


Laboratory or continuous (for washing loss calculation)

Table 1: Process measurements required for DD washer control

The washer controls use commonly available process measurements including the pulp feeding pressure, the pulp pressure difference in the screen unit, the washing liquor feed pressure, the drum rotation speed and the return liquor conductivity. Conductivity is an indirect measurement that indicates washing efficiency by inferring the amount of dissolved solids.

In addition to this normal process instrumentation a number of soft sensors have been developed to model the washing process. These soft sensor are calculated or deduced values which use other process measurements as inputs to their calculation. These soft sensors include:

  • Delayed conductivity
  • Delayed consistency
  • Rotation speed target
  • Conductivity target
  • Partition consistency

The most important soft sensor is pulp consistency calculated by mass balance or by inference from the load applied on the scraper motor, for example the bottom scraper of the digester or the top scraper of the diffusion washer. The second most important soft sensor indicates washing loss by combining a laboratory measurement of COD and a continuous conductivity measurement. The COD lab measurement provides an absolute baseline while the online sensor provides a fast-responding real time measurement. These discrete and continuos measurements are combined in a product called D2C.

In this type of rotating drum washer the drum rotation speed is the directly controlled variable. The retention time, which is a factor in the Washing Potential in Diffusions model, is proportional to the drum rotation speed. The slower the drum the longer the retention time.

In the DD washer optimization program the drum rotation speed is kept as low as possible to maximize the pulp washing efficiency. The rotation speed is controlled using fuzzy logic. The control cannot set a rotation speed target that is too low as the pulp consistency may become too high and the washer may become stuck. In normal operation, the rotation speed is adapted to production rate. In the DD washer, the feed pressure is controlled by rotation speed.


The improved stability of a washing line is demonstrated by the better control over filtrate tank levels. This stable operation is particularly important during wood species changes and production rate changes. Figure 5, 6, 7 and 8 show how optimization controls can stabilize filtrate tank levels in a diffusion washing line. The washing line configuration for this specific case is shown in Figure 9.

Figure 5 shows the trends of the first diffuser filtrate tank level and Figure 6 shows the level of the third diffuser filtrate tank level. In both cases, the trends are before optimization. For comparison, Figures 7 and 8 show the same trends after optimization. The times of wood species changes are indicated on each of the graphs.

Figures 5 & 6

Figures 7 & 8

Figures 5 and 6 show the filtrate tank level for the first diffuser filtrate tank (Figure 5) and for the third diffuser tank (Figure 6) before optimization controls were implemented. The same trends after optimization controls were implemented are shown below in Figures 7 and 8 respectively.


All these optimizing controls need to run without operator intervention to be effective. The controls obviously must work smoothly and avoid process upsets. The control interface is through a single page in the control system's user interface. Optimizing controls are activated through a single command on this page. See Figure 9.

Figure 9

Figure 9: An example of an operator interface display


Washing line optimization has been implemented at four pulp mills in Finland. Figure 10 shows the results of one mill where COD washing losses were reduced by about 25% by more efficient washing. The outlet pulp consistency also rose by 0.6%.

This mill reports that the washing load has been distributed in an optimum way between the washers in the line The line is now less susceptible to disturbances. There have been benefits in other related mill departments. Bleaching chemical consumption is lower. The dry solids content in the weak liquor to the evaporation plant is 0.5% higher and the variability is 30% less.

Figure 10

Figure 10: Brown stock washing optimization results from a kraft pulp mill in Finland

At another mill which produces hardwood and softwood pulp with species changes every 72 hours the washing losses before oxygen delignification and after have been reduced by about 25%. The results of both hardwood and softwood production runs are similar. See Figures 11, 12, 13 and 14.

Figures 11 - 14

Figures 11, 12,  13 and 14. Washing losses have been reduced before oxygen delignification and before the bleaching operation on both softwood and hardwood grades.


A new method to optimize the brown stock washing line has been developed. This method is based on the optimization of individual washers and then the whole washer line. Washer optimization is made possible by a new process modeling technique called Washing Potential in Diffusion.

By the implementation of washer-specific models and optimization controls the stability of a brown stock washing line can be improved significantly. Optimization makes operator's work easier and eliminates human errors. With these controls, the entire line is more tolerant of disturbances and recovers from abnormal situations more quickly and effectively.

With this improved stability the overall washing efficiency of the line can be optimized to reduce washing losses and increase filtrate solids. This can result in reduced bleaching costs and evaporation costs. Improved pulp and filtrate quality can also help to eliminate production bottlenecks.


1. Crotogino, R.H., Poirier, N.A., and Trinh, D.T. The Principles of Pulp Washing. Tappi Journal, June 1987, pp 95-103.

2. Ulvinen, Mikko. Optimizing Controls for a Brown Stock Washing Line. Master's Degree Thesis, Tampere University of Technology,  Tampere, Finland. January 1999.

3. Johan Gullichsen and Carl-Johan Fogelholm, Chemical Pulping, Fapet Oy. Jyväskylä, Finland 2000.


[Home] [Title] [Author] [Organisation] [Keywords]