Holik H. (ed.) Handbook of Paper and Board

Подождите немного. Документ загружается.

be stopped by the MCS in less than 20 ms. The cycle time of a MCS is therefore

in the 5 to 20 ms range.

• DCS, distributed control system:

– Analog process controls

Analog process controls are mainly simple standard controllers. Examples are

level controllers in chests, pressure controllers in pipes, etc. Most process con-

trollers act much more slowly than the MCS controls. Actions happen in sec-

onds to minutes. The cycle time of typical DCS functions is therefore in the 1 to

4 s range.

– Wet end controls

Some controls in the process are more complex than just PID control loops. One

example is the wet end control which ensures uniform process conditions in the

approach flow of the paper machine and in the forming section by maintaining

constant consistencies, gas content, charge and other parameters. It involves

e.g. special software tools (soft sensors) using wet end data instead of direct

measurements to predict basis weight during start up of the machine, when the

paper has not yet reached the reel or the quality measurements. Such, wet end

controls not only reduce the start up time of the process, but also minimize the

consumption of chemicals, i.e. only as much retention aid is used as is really

needed to reach a required retention level in the forming section.

• QCS, quality control system:

– Paper quality measurement

The QCS has on-line sensors for basis weight, ash content, moisture content,

gloss and other paper quality parameters. The sensors are mounted on measure-

ment frames, which allow the sensors to move (scan) within 20 to 40 s across the

paper width. These sensors measure the paper quality parameters in cross ma-

chine direction (CD) and machine direction (MD). Each scan typically gives

about 1000 measurement values (databoxes), equally spaced across the web

width. An array of CD measurement values is also called a CD quality profile or

CD profile.

The average measurement of each scan gives one MD quality value. A time

series of MD measurement values is also called an MD quality profile or MD

profile.

MD measurement values and CD profile measurements are available only every

20 to 30 s.

– Paper quality control in the machine direction

The first target of quality control in MD is to keep the main quality parameters of

the produced paper within the limits defined for the current production. As an

example, basis weight control is achieved by continuously adjusting the stock

valve, which feeds the blended stock to the machine.

The second target is to ensure fast grade changes, i. e. to ensure a fast change of

basis weight, moisture and ash content to meet the specification of the next

produced grade with a short transition time.

The third target is to allow the operator to change machine settings without too

large an impact on paper quality. The most important example is the coordinated

9 Control Systems for Paper Machines402

speed change control. This is a tool to allow machine speed changes without

disturbing paper quality during the transition period. The effect of a speed

change is compensated by coordinated adjustments of stock valve, ash valve and

drying energy, based on a dynamic process model.

MD controls are slow but complex control loops. Only two to three measure-

ments are available per minute. Using interpolation techniques or Kalman fil-

ters, control steps can be performed every 5 to 10 s. MD controls are complex

because of the considerably large transport dead time between actuators (i.e.

stock valve) and measurement (i.e. scanner at the reel), which is typically larger

than 1 min.

Additionally the quality parameters have some influence on each other. Opening

the ash valve increases not only the ash content but also the basis weight and

moisture. Increasing the basis weight leads to higher moisture content at the

reel, if the drying energy is not increased simultaneously. Thus, a decoupling

controller with dead time compensation is required.

– Paper quality control in cross machine direction

Many paper properties can be controlled in small control zones across the ma-

chine, to ensure uniform quality in CD. Such actuators are located e. g. in the

headbox to control basis weight, in the press and in the dryer section to control

moisture content, and in the calender section to control caliper across the paper

width.

CD controls act even more slowly than MD controls: The measurement noise on

zonal measurements is much higher than the noise on MD measurements. The

measurement time of one zonal measurement (databox) is around 20 ms,

whereas MD measurements are averages of typically 1000 databoxes (1 scan).

The special complexity of CD controls is the coupling of adjacent control zones.

Each control zone affects the quality of some neighboring control zones. The

controller needs to know exactly the spatial response of actuator movements in

the CD. On the other hand the exact location of the response of a CD actuator in

the paper may change over time e.g. due to changing shrinkage of the paper in

the drying section, or by web wander to the left or right.

– Paper quality reports

The QCS not only controls quality, but also documents the achieved quality,

which is required according to ISO 9000 and other quality regulations.

The quality reports also serve internally within the paper mill as a means of

supervising the efficiency of the production process, by reporting per week, per

day or per shift how much paper of which quality has been produced.

• MCC, motor control center:

– From these cabinets hundreds of motors on the machine are controlled.

These are motors for pumps, movements of machine parts or others. The MCC

is additionally used to control a few other consumers of electrical power, for

example heating devices for oil in calenders.

– The MCC is not responsible for the roll drives in the paper machine. The

drive system is usually an independent package, which is delivered by the ma-

chine drive supplier, and which receives its setpoints from the DCS.

9.1 Objective and General Terms of PM Control Systems 403

• Monitoring systems:

– Bearing and vibration monitoring

Acceleration sensors are attached to bearings and close to roll shafts, to detect

bearings at risk of failing and roll roundness abnormalities before they have a

negative impact on the process. Thus bearing and roll changes can be done

during the next planned shutdown of the machine, and do not cause extra

downtime.

– Monitoring of periodic paper quality variations

Typically 30 to 100 process data are measured in parallel. The measured signals

include important quality data, taken from the QCS sensors. The target is to

detect the occurrence and the reasons for periodic paper quality variations.

– Web break monitoring

Web break monitoring is a camera-based tool to find the reason for paper breaks,

and as a consequence to improve efficiency of the production process.

– Web inspection system

Web inspection systems are also camera based. Cameras are mounted in the CD

direction across the web, and check the complete production for defects like

holes, dark spots, white spots.

– Information system:

An information system mainly fulfils three tasks:

1. It collects on-line data of MCS, DCS, QCS and monitoring systems into a

central data storage.

2. It provides a user interface to view the historical data of these systems, allow-

ing one to compare and correlate the data from the various origins. The most

convenient user interface is based on web browser technology, similar to the

Internet. Some of the data are presented to the user as time trends, others as

statistical reports over a defined production period.

3. It provides tools for data analysis. This usually includes the capability of show-

ing the fourier transform or the cross correlation of selected data. More sophisti-

cated information systems contain expert systems or data mining tools, to auto-

matically detect relationships between data.

• Automation system:

All the systems above together form the complete automation system of a paper

machine (Fig. 9.1). The automation systems are grouped around a fiber optical

data network, which has a ring topology for redundancy reasons. The connected

systems are operating stations, engineering stations for configuration or main-

tenance work, DCS cabinets, server computers and dedicated control computers

for various monitoring and control purposes. The dashed lines indicate the dif-

ferent locations with automation equipment, such as control rooms for opera-

tors, electrical control rooms (rack rooms) with DCS or QCS computers, or the

paper machine.

9 Control Systems for Paper Machines404

Fig. 9.1 Example of an automation system – overview (source: Voith).

9.1 Objective and General Terms of PM Control Systems 405

To summarize, nowadays systems can be divided into the following categories

• DCS/MCS

– Control systems for analog and digital controls

– Capable of implementation of advanced model based controls which are

needed for e.g. soft sensor or QCS applications

• QCS

– Quality measurements for paper properties (direct measurements)

– Measurements for felt and wire condition

– Indirect measurements of paper properties, also called soft sensors

– Quality controls (sometimes integrated into DCS systems)

• Monitoring systems

– Bearing and vibration monitoring

– Monitoring of periodic paper quality variations

– Web break monitoring

• Information system

– Process history database

– User interface

– Diagnostic tools

• Automation system

– Includes all of the above systems

In the following chapters some elements of a paper machine automation system

are discussed in more detail. The focus is on QCS and information systems, as

they are mainly responsible for the ability of the mill to maximize quality tons on

the reel.

9.2

Quality Control System (QCS)

9.2.1

Quality Measurements

9.2.1.1 Scanning Measurement

Most paper quality measurements in the paper machine are scanning measure-

ments. In various locations so called “scanners” with usually one upper and one

lower measurement head are found (Fig. 9.2). Both measurement heads move

synchronously across the paper width. One scan takes about 20 to 30 s. The paper

moves much faster than the scanning measurement. While the paper is moving

e.g. 500 m the scanner moves across the web only 7 to 10 m. Thus, the scanning

sensors can be seen as a system which measures paper quality in MD and which is

moved slowly across the web.

During each scan hundreds of measurement values are collected. The resolution

in CD is typically 10 mm. Each scan contains information of MD and CD variation

of the measured paper quality. MD and CD information are used for different

purposes and have to be separated from each other.

9 Control Systems for Paper Machines406

• The stable CD profile is given by a weighted average of multiple scans. Typically

the displayed CD profiles on an operator screen correspond to an average of

9 scans. Averaging is needed to suppress the MD variation and the residual

variation of the process. Thus a CD profile contains the average of the last 3 to

5 min.

• The MD profile is the time series of average values of the individual scans.

Advanced methods use e.g. Kalman filters to generate a best estimate of the

current MD value every 5 s, by assuming, that the CD profile has less dynamics

than the MD profile.

9.2.1.2 Fixed Point Measurement

Besides scanning measurements there are also a few fixed point measurements

for paper properties which are assumed to be constant across the web width. For

example color measurements are sometimes fixed point measurements as color

should not vary across the web width.

9.2.1.3 Basis Weight

Basis weight is measured in g m

–2

and it is the total mass of 1 m

of paper, includ-

ing all components, like fibers, fillers and water. The sensors consist of a radio-

active source and a radiation detector on the opposite side of the paper. The absorp-

tion of radioactive radiation is a measure of the total mass of the paper. Radioactive

sources are usually promethium, krypton or strontium. Promethium provides the

signal with the highest sensitivity and can be used for basis weights up to

250gm

–2

. Strontium is used only in very heavy board applications.

Nowadays detectors are either ion chambers or solid state detectors. Ion cham-

bers are gas filled tubes and require a high voltage to detect electrons. Solid state

detectors use the photovoltaic effect to detect electrons.

There are also other possible ways to measure basis weight, such as X-ray tubes

to generate radiation, or spectroscopic methods, but currently they cannot compete

with the sensors based on radioactive radiation.

Fig. 9.2 Quality measurement – scanner (source: Voith).

9.2 Quality Control System (QCS)

407

The accuracy of a basis weight measurement is typically 0.1 g m

–2

. To get a

better understanding what this figure really means, it is good to know that the

surface of the detector is smaller than 10 cm

. Thus, if a sample is fixed between

the sensor heads with a size of 10 cm

, the weight of this sample is measured with

a precision of 0.0001 g.

9.2.1.4 Moisture

Moisture is measured in per cent of basis weight. The measurement is usually

performed by analysing absorption of infrared light at three to four different wave-

lengths. Two of the wavelengths match the absorption peaks of water and fiber.

The other wavelengths are taken for reference purposes. Modern moisture meas-

urements have internally four detectors to make sure that all wavelengths are

measured simultaneously, whereas older measurements used only one detector

and a spinning filter wheel.

For high weights above 300 g m

–2

infrared radiation can no longer pass the

sheet. Therefore in this case microwave radiation is used instead. The measured

physical effect is that moisture slows down the speed of microwaves, which can be

detected by e.g. evaluating the phase shift of the signal.

The best achievable accuracy of moisture sensors in a PM today is 0.25% relative

moisture content.

9.2.1.5 Fillers

Fillers or “ash” mean the amount of inorganic material within the paper. The

measurement output is either g m

–2

or % of basis weight. Typical inorganic com-

ponents are:

• clay

• calcium carbonate (CaCO

)

• titanium dioxide (TiO

)

Filler content is measured by absorption of radioactive radiation or by absorption

of X-ray radiation. Thus the principle of filler measurement is similar to basis

weight.

Radioactive radiation is monochromatic and stable. The radiation has one ex-

actly defined wavelength. Clay, CaCO

and TiO

have different absorption coeffi-

cients for radiation of the given wavelength. Thus,

• for measurement with radioactive radiation, the percentage of each filler compo-

nent has to be known beforehand. With deinked pulp (DIP) an additional meas-

urement of the ratio of the different components is required. This can be carried

out by using XRF (X-ray fluorescence)

• for X-ray radiation the spectrum of the radiation source has to be tuned by filters

to compensate for the different absorption characteristics of the different fill-

ers.

9 Control Systems for Paper Machines408

The accuracy of ash measurement today is typically 0.1 g m

–2

, similar to basis

weight.

9.2.1.6 Caliper

Caliper is measured in micrometers. There are three main ways to measure cal-

iper:

1. Contacting caliper measurement: Two measurement fingers contact the paper

from both sides. The distance between the fingers is measured using electromag-

netic fields. This method is used when accuracy demands are high. For high speed

paper machines the contacting measurement has many disadvantages:

• At the contacting point the paper becomes polished. A small glossy stripe may

be visible on the paper surface.

• If the stock is not of best quality and if for example a small hard or sticky piece

happens to be at the paper surface, the contacting sensor will tear it out. This

may result in a hole in the paper or even in a web break.

• The fingers are subject to wear as minerals in the paper grind the measurement

fingers when passing by at high speed.

The accuracy of this method depends on the application and is between 1 and

0.5 microns. The relative accuracy, which is needed to measure the shape of a cross

profile, can be better than 0.25 microns.

2. Air bearing caliper measurement: The sensors are mounted onto flat meas-

urement plates with a diameter of roughly 60 mm. Those plates are drilled, with

pressurized air coming out of the tiny holes. The plates are pressed towards the

paper from both sides. The distance of the plates from the paper is controlled by

the air flow. The distance between the plates is measured using an electromagnetic

field.

The thickness of the paper is calculated by subtracting the thickness of the air

bearings from the distance of the sensor plates. As the thickness of the air bearings

is not known precisely, this method is mainly used as a relative measurement to

detect MD and CD profile variations of very sensitive papers, where a contacting

measurement is not applicable. But also the relative accuracy is much less than the

accuracy of the contacting sensor.

3. Optical caliper measurement: Due to the deficiencies of the air bearing sensor

and the need for a precise noncontacting caliper measurement, most sensor man-

ufacturers currently push the development of optical noncontacting measurement

methods. They consist of three sensors:

• An electromagnetic distance measurement which measures the distance be-

tween the two sensor plates on opposite sides of the paper.

• Two optical distance measurements which measure the distance between each

sensor plate and the paper.

The difficulty here is, that the paper surface is optically not well defined. The outer

layer of the paper is less dense or may even be transparent. Thus optical measure-

9.2 Quality Control System (QCS) 409

ments “see” the paper surface as being about 5 microns beneath the real physical

surface. Therefore, the accuracy of the existing optical measurements is for most

applications still not sufficient.

9.2.1.7 Coat Weight

Coat weight measurement usually requires two measurements, one before and

one after the coater, which are subtracted from each other. These measurements

can be basis weight measurements or ash measurements. Coat weight is meas-

ured in g m

–2

Nowadays in some cases single sided infrared coat weight measurements are

used, as both sides of the paper are coated simultaneously, and the paper maker

wants to know the amount of coat on each side. Such infrared coat weight meas-

urements are hard to calibrate, as the infrared beam penetrates through the coat-

ing color into the paper, and the paper also contains minerals and pigments. Thus

the accuracy of this kind of coat weight measurement varies a lot, depending on

the application.

9.2.1.8 Color

Color measurement is a collection of optical measurements. The basic measure-

ment is color itself and is presented to the user in a color coordinate system. A

common color coordinate system is for example CIELAB, where the parameters a*

and b* define the color space and L* defines the brightness.

Paper color is usually defined as the reflection of the surface of a pile of paper in

defined light conditions, where all paper sheets in the pile have the same color.

Thus the following difficulties arise in an on-line measurement:

• In a paper machine there is a running web, i.e. one single layer of paper. The

background of the paper is visible. Thus, the on-line measurement has to com-

pensate for the opacity of the paper. This is performed by two measurements in

sequence, one with a white background, one with a black background.

• Fluorescence transforms UV light into visible light. Thus the fluorescence effect

changes the appearance of the paper e.g. in daylight. Fluorescence has to be

measured separately when optical brighteners are added to the paper.

This explains why brightness and opacity measurements are usually part of a color

measurement.

9.2.1.9 Gloss

Gloss is a measure of the reflectance of the paper at a given angle. The measure-

ment is calibrated:

• The appearance of black velvet corresponds to the measurement value, i.e. a

gloss value of zero.

• The appearance of black glass corresponds to the gloss measurement value of

100%.

9 Control Systems for Paper Machines410

9.2.1.10 Others

Other available sensors include the measurement of formation, roughness, and

porosity.

Further sensors are still in the process of development and cannot be treated as

standard sensors. These include sensors for fiber orientation, strength properties,

etc.

However, there still remain many important paper properties which are not yet

measurable on-line, and where on-line measurements will probably not be availa-

ble in the near future. Most important to note in this category is printability. A

printability sensor, e. g., would be able to predict “missing dots” in a real printing

test.

Some of those properties will be measurable using so called “soft sensors”. For

example a soft sensor for porosity would “measure” porosity based on a calibration

of about 100 process data against corresponding laboratory measurements of po-

rosity.

9.2.2

Quality Control

9.2.2.1 Machine Direction Control

Machine direction control uses MD measurement values to control the down-

stream quality of the paper. Advanced MD controls are based on physical process

models of the papermaking process. This means, mathematical models describe

the physical process and are able to predict the effect of control actions on the

process and on paper quality. Such model-based controls are especially essential to

achieve

• Start up control: to control the paper quality as fast as possible during start up of

the machine. This is a topic for machines which restart quite often during the

day, as for example off-machine coaters.

• Coordinated speed change control: to keep quality constant even during speed

changes. This requires a coordinated change of at least stock, ash, drying ca-

pacity and speed. For example: If the stock valve is turned it takes a while until

the consistency change arrives at the headbox. This is due to the transport delay

time between the stock valve and the headbox, and also to the mixing time in the

chests. The increase in machine speed should start when the consistency at the

headbox changes (Fig. 9.3). Control of ash content and the settings of the dryer

section have to be done in a similar way.

• Production maximization control: to increase the machine speed when the proc-

ess is not at the limit, for instance when drying capacity is still left, and machine

speed is below a user defined maximum speed. In fact this involves a series of

coordinated speed changes.

• Grade change control: to keep the time for a grade change as short as possi-

ble.

9.2 Quality Control System (QCS) 411