[Home] [APPW 2004] [Journal papers]




Andy Worsick and Stephen Halsall


Sandusky Walmsley Limited


efficiency, optimisation, losses, cleanliness, stability, process control, draws, threading, discipline






Sandusky Walmsley have carried out efficiency and optimisation studies on a number of paper and board machines around the world.

A methodology for undertaking these studies, which involves first identifying and then following the losses, is briefly outlined.

A number of the causes of efficiency that have been identified in these studies are common at many mills. The five that occur most frequently are: -

  • Wet-end and Press Cleanliness
  • Process Control Swings
  • Press Draws
  • Threading
  • Post-Reel Losses

These are detailed with the use of case studies to explain the problems. Although not all may be present in every mill, it is our experience that a number of them invariably are.

Finally, technologies and techniques for overcoming these five causes of loss of efficiency are explained. In some cases, the solution has negligible cost, whereas in others a small capital investment can have a large impact.

The paper concludes with an illustration of the effect that improving efficiency has on runnability and profitability.


The first step in improving paper machine efficiency and runnability is to identify the major causes of inefficiency. To do this, we developed a "Follow the Losses" methodology to assist in carrying out efficiency studies.

The technique that we have developed is based on an initial analysis of lost tons and lost time. Availability of this data is a key factor in undertaking an efficiency study, and provides the starting point for any investigation.

The data are typically divided into a number of areas: -

  • Breaks
  • Unplanned downtime
  • Planned downtime
  • Post reel losses

Values for a particular machine are benchmarked against other similar machines, and against what would be expected of a machine achieving good or even world class standards. Expected values for average and good machines are shown on Fig 1.

Figure 1

Figure 1

This initial data analysis will highlight areas that should be investigated during an efficiency study. It is important to concentrate the study on those areas where improvements can be expected, and where those improvements will have a significant effect on runnability and profitability.

Having carried out a number of efficiency studies in recent years, we have identified the most commonly occurring causes of efficiency loss. Whilst it would be unusual for all to occur on a single machine, it is not uncommon to see two or three of these in most mills.

The five most common causes of efficiency loss (with some Case Study examples) are detailed below:



When analysing break data, it is common to see a high number of wet-end and press breaks listed. Often, these are a result of poor cleaning practices. Fibres and dirt can build-up in many places around the wet-end and press. It is important to identify which areas of build-up are important to address. The key is to look for any place where the fibre can drop off onto the sheet, or where it may drop onto the inside of the wire loop - particularly in areas close to the headbox (e.g. breast roll doctor in Fig 2). In the press, areas inside the felt loop on the ingoing side of the press nips, or areas directly above the sheet, are of most importance.

Figure 2

Figure 2

These fibres can be pressed onto the sheet, or through the wire, and will often result in a weak spot in the sheet that will cause a break as it travels down the machine. Camera systems are installed on many modern machines, and can often be used to confirm the cause of a break. On machines without camera systems, analysis of the broken sheet on the reel can also be used. For example, on bleached grades, a build up of white fibre at the break point normally suggests a wet-end dirt problem. If the fibre is brown, it is more likely to come from the press.

It should also be remembered that build-up that is outside the sheet width, or not inside the wire loop, is much less of a problem. It is not uncommon to see operators carefully cleaning handrails and platforms, whilst ignoring build-up on the under side of a wet end saveall for example.


A number of recommendations have been made to address this issue.

  • Operator training - often just making the operators aware of the important areas that need to be cleaned regularly can make a big difference.
  • Additional cleaning hoses - it is very difficult to thoroughly clean some of today's wider machines from the front-side only. Adding high-pressure hose reels at the backside of the machine can give the operators an important tool to help maintain good cleanliness.
  • Cleaning regime - don't try to clean every point at every break. A diagram of the most important areas (see Fig 3 for a typical example) can be displayed in the operator control room. At any one break, the operators should just clean two or three areas, and note these in the shift log. At the next break, the operators should continue down the list. Thus, all areas will be cleaned after four to five breaks.

Figure 3

Figure 3



It is extremely difficult to operate a machine efficiently when there is any variability into the stock flow or services to the paper machine. Changes in furnish characteristics can contribute to breaks on the machine, or result in quality losses.

Variables that have previously been found to be out of control include: -

  • Retention aid addition rate
  • Broke addition rate
  • Broke consistency
  • Machine chest consistency

On one board machine, the wet web was observed to be tightening and then slackening at the couch to press draw, and between the first and second presses (Fig 4).

Figure 4

Figure 4


Operators must carefully monitor the key variables of their incoming furnish. Feedback to the pulp mill, or the stock preparation area, is required. All process control loops must operate effectively, and any transmitters / actuators must be in proper working order.

In other cases the cause of the variation must be carefully investigated. In the example given above initial thoughts were that the problem might have been a result of drive instability. This was checked, but appeared to be running in a steady state. The headbox pressure control was level, and no variations could be seen in the fan pump operation (amps and rpm). Back-tracking further upstream, we eventually found that the level in the thick stock headbox (stuffbox) was oscillating. It was varying by 0.2m, with the head over the valve being 4.0m. This results in a 5% variation in basis weight, and a similar moisture variation, which was contributing to the high number of breaks on the machine.



Many modern paper machines operate with closed draws from the former to the press, and between the press nips. In some cases, the draws are closed within the dryer section as well.

On machines where an open draw exists in the press section, it is common to experience a high number of breaks at this point. The first point at which the relatively weak wet web is unsupported, and subject to tension forces at the draw point, is a classic opportunity for the sheet to break. The longer this draw is, the worse the problem will be.

Fig 5 shows an open draw between a 3rd and 4th press - the unsupported sheet length is currently around 150mm.

Figure 5

Figure 5


Rebuilds to close up the press draw are relatively simple and low-cost in many cases. In the case above, the open draw length should be reduced to 15 - 25mm.

Fig 6 shows a rebuild to a Cluster Press + 3rd Press combination, where the draws between the 2nd and 3rd press, and between the 3rd press and the dryer section have been reduced.

Figure 6

Figure 6



As well as reducing the number of sheet breaks, efficiency can also be improved by reducing threading time. Threading time has a direct impact on total non-papermaking time. It can also affect the paper machine's subsequent performance, because some rolls might heat up, the furnish might change, and the scanner has nothing to scan. In all these cases, the shorter the threading time, the quicker the recovery of the machine systems.

In many mills threading is quick and efficient. Unfortunately, however, in many mills it is not. In our analyses of mills we have seen a wide variety of equipment and techniques, and a correspondingly wide variety of problems and delays. There is no single answer, but three of the more common issues are set out below.


In a typical survey week we might watch and record five thread-ups. It is not uncommon to see threading times vary from 5 to 25 minutes, with a different problem each time. This suggests that the systems being used, whether mechanical or human, are not repeatable to any high degree. Take a rope system as an example. The tail may go all the way through, or fall out in one of several places. This could be simply due to the way the tail is put into the ropes. If the system has repeatability, the tail should sit properly in the ropes each time. If the system is haphazard, the tail could be twisted in the ropes, or maybe just held by the tiniest corner. Repeatability is the key to understanding and correcting problems. In general, mechanical systems are more repeatable than hand systems, and fully automatic systems are more repeatable still.


One of the more irritating causes of threading delay is the readiness, or lack of it, of a downstream process. We have seen on many occasions a tail get through to a machine section that is not running, simply because no-one had restarted it after the break. We have seen operators clearing up the floor while the tail sat on the last dryer. We have seen a tail through onto the reel doctor before anyone had looked for a clean reel spool to turn up onto. We have many times seen unexplained pauses of two or three minutes between threading operations. Crews are most often extremely busy during threading, but not on the important tasks.

The keys to short threading times are a plan and a degree of urgency. Not panic. Machine operators should be aware of specific tasks to be done before threading is attempted, and of the prescribed sequence of actions that get the tail to the reel fastest. They must also be focussed on the job in hand, which is to get the machine making paper again.


Most machines use ropes to thread some sections of the machine. Rope systems have traditionally been the most altered, adjusted, tinkered-with and altogether ruined component on the machine, which is unfortunate given their importance in machine efficiency. Pulleys and rope-runs are moved to fit pipes in, to open up walkways, to allow broke clearance, and sometimes even to improve threading efficiency. The result is often poor threading and poor rope life.

In one mill, we were asked to fix a rope system that we had installed some years previously. We suggested, and supplied, a system identical in every respect to that which had been supplied originally. It was installed and worked well. One year later, we were asked to do exactly the same again. Good for suppliers; bad for mill efficiency.

Often with little investment, rope systems can be improved and efficiency increased by a review of the pulley positions, particularly at transfers.


The several threading problems we describe above, and probably several more, are best addressed by a thorough study of the threading process. This study should aim to do two things. Firstly, to identify those areas of the threading process where the equipment is not helping. Secondly, to identify inefficiencies in the activities of those doing the threading. This latter aspect goes somewhat beyond the remit of a paper machinery builder, but it is a critical aspect to high efficiency operation.


Problem One

All paper machines make more paper than is needed on each jumbo reel and the excess has to be slabbed off as broke. The key to high efficiency is keeping this slab loss to an absolute minimum.

At one mill, the sales department had standardised on only two different shipping roll diameters to suit the needs of their customers. The machine aimed to produce sufficient paper on each jumbo reel for either 3 sets of 1250mm diameter shipping rolls off the winder, or 5 sets at 1000mm diameter. These values equate to jumbo roll weights of approximately 17,300kg and 18,700kg respectively.

Fig 7 shows the actual weights (arranged in ascending order) of over 150 jumbo rolls analysed. It was expected that two distinct bands of reel weights would be observed (shown by the shaded areas on the graph). The lower band would correspond to the weight of individual jumbo rolls produced when making 3 sets of 1250mm diameter shipping rolls, and the higher band to 5 sets of 1000mm diameter shipping rolls. Every point outside these bands represents a jumbo with an excessive remaining slab.

Figure 7

Figure 7

The analysis actually showed a massive variation in roll weight, to the extent that the data plotted is almost linear. This explained the slab loss, which at the time was running at over 10%.

In another mill, a detailed analysis of slab loss on several machines showed great inconsistency, and a varying set requirement (to meet customer order requirements) that caused more upset to the process.


In many cases, operator discipline is the key factor in controlling post-reel length losses. Careful attention to production planning, which will assist the operators in consistent operation will also give improvements.

One recommendation has been to suggest appointing someone with the specific duty of monitoring and improving length loss for a period of a few months. This person needs to analyse length loss on a daily basis, and to observe and challenge normal practices.

Many modern machines have some system of paper length measurement prior to the reel. This length measurement can be correlated with the requirements for each jumbo (and subsequent number of sets and shipping roll diameters), and used to give a visible or audible warning to tell the operators to instigate a turn-up. Again, discipline here is key. At one mill, when the turn-up alarm light flashed, the operator slowly walked towards the reel bench board, made his safety checks, including making sure no-one was in the vicinity of the reel, checked the new spool was running at the required speed, and then pressed the turn-up button. This whole process took around 30 seconds, which no doubt to the operator seemed insignificant. The reality is that he had just scrapped 1.25% of the machines production (based on a 40 minute jumbo build time), which as will be seen later, had a large effect on his employer's profitability.

The recommendation in this case was to use the length measurement system to trigger a two minute warning, as well as the actual turn-up alarm.

Problem Two

Quality losses after the reel are another common cause of efficiency loss. On older reels, it is not uncommon for over 25mm thickness of paper to be left on the spool at the end of the winding process, which is typically slabbed off into the dry end pulper.

The reason for leaving this amount of paper is often the occurrence of crepe wrinkles (on a newsprint sheet for example), or other sheet defects at around 15 - 25mm from the spool.

A jumbo roll structure is required that will allow the operators to unwind the roll most efficiently in the winding operation.

The structure required in order to reduce losses close to the spool, and to prevent any defects from occurring later in the roll is to start with a tightly wound roll. The wound-in tension should then be reduced as the diameter increases, so that the stresses generated by the outer layers of paper never exceed the ability of the inner layers to support them.

With conventional reels, this structure is not achieved. The nip pressure between the spool and the reel drum during a turn-up is the sum of spool weight, hook weight, and hook loading pressure. The spool weight and hook weight components are variable depending on the position of the primary arms.

Most reels use a level rail to support the winding roll during the majority of the roll build-up. A pair of secondary arms that pivot from near the floor provide the loading force that holds the building roll against the reel drum. A constant pressure in the loading cylinders results in a non-constant nip pressure between the roll and the reel drum due to the changing loading geometry as the roll increases in diameter.


To achieve control of roll structure during turn-up, while the spool is held in the primary arms requires the addition of primary arm nip relieving. This consists of an additional loading cylinder mounted in the primary arm that acts to lift up on the spool bearing housings. Enough force is applied with this air cylinder to offset the nip forces generated by the spool weight and hook weight. The hook loading pressure then determines the nip pressure between the spool and the reel drum.

To sense the position of the primary arms so that the nip relieving pressure can be maintained at the correct value to offset the spool and hook weights an electronic position transducer is mounted to the end of the primary arm cross-shaft. This transducer sends a signal to a PLC, which converts the signal to primary arm position.

To obtain the desired roll structure control while the roll is being held in the secondary arms, requires the ability to sense the roll diameter and control the loading pressure in the secondary arm cylinders.

Secondary arm programmed loading requires an electronic position transducer mounted to the front secondary arm pivot pin that senses the angular position of the secondary arms. This transducer sends a signal to the PLC, which controls the loading pressure in the secondary arm cylinders.

With a conventional reel design, the jumbo starts to be wound with the primary arms in the vertical position. These arms then transfer to the horizontal position, applying a load to the spool throughout. When the primary arms reach the horizontal position the secondary arms move in and also load the spool. It is only when the secondary arms have engaged that the primary arms release and move back to the vertical position. This overlap period, giving rise to a 'bump' in the nip pressure profile, causes a high load of the reel against the reel drum for a finite period of time (Fig 8). This high load causes a hard ring of paper, which compresses the layers of paper beneath it and allows the tension in these layers to relax.

Figure 8

Figure 8

In order to eliminate this effect, "bumpless transfer" systems are recommended. The control of primary and secondary arms at the transfer stage is carefully designed. The primary arm loading is decreased as soon as the secondary arms start to load, thereby ensuring a smooth transfer (Fig 9).

Figure 9

Figure 9


The first step in addressing paper machine inefficiency is to collect data detailing the major areas of loss.

From this point, time and effort can be spent on identifying the major causes of efficiency loss.

Once identified, there are often low (or even zero) cost solutions to address these causes of inefficiency, which will have a significant effect on the runnability and profitability of a paper machine.

For example, a 9-metre trim newsprint machine, operating at 1400 m/min on 48.8gsm newsprint, will produce 323,000 Tonnes/Year at 100% efficiency.

A one percentage point increase in efficiency will result in over 3,000 Tonnes/Year additional saleable production.

Assuming a marginal contribution of 150 Euros per ton, this results in almost 500,000 Euros being added to a machine's profitability for each 1% improvement in efficiency.


The authors would like to acknowledge the assistance and encouragement given by Mr Ian Binns, Sandusky Walmsley Limited - Technical Consultant.


Whilst examples from various machines have been used to illustrate the case study examples given above, the source of this information is confidential, and details of individual machines will not be disclosed.


Andy Worsick (Co-author and presenter)

Andy Worsick graduated from the Paper Science Department at the University of Manchester Institute of Science and Technology (UMIST) in 1989, with a Bachelor of Science Degree.

On completion of his studies, he joined Beloit Walmsley Limited in Bolton, England, where he worked as an Applications Engineer, and then Product Manager, specialising in Multi-ply Board Grades.

Six years with Beloit were followed by a spell at Simon Holder Ltd (a smaller UK paper machine builder), where he worked as a Sales Manager covering paper mills in the North of England.

In November 2000, he returned to Sandusky Walmsley (as the company is now known), where he currently holds the position of Sales and Applications Manager.

Andy works on the design and specification of Capital and Rebuild projects for paper and board machines, and travels extensively to mills around the world.

He has presented papers at two previous TAPPSA conferences, as well as at APPITA (Australia), PAPTAC (Canada) and PITA (UK).

Steve Halsall (Co-author)

Steve Halsall is a graduate of the Mechanical Engineering Department of the University of Bradford. He joined Beloit Walmsley in 1992.

Steve's early responsibility was the design and engineering of paper machine dryer sections. This was followed by a year in the Analysis and Optimisation Department. In 1998 he joined the Sales Department

Steve is currently Sales and Application Manager within the Sales Department, where his responsibilities include technical specification and sales of many large paper machine projects. He also takes part in several machine surveys each year.