I Shoe nip calendars comprise one or more roll nips, in which the fibre web is calendered while passing through the roll nips. The roll nips are formed between the shoe nip and the opposite heated thermo roll. One of the most commonly used shoe roll types comprises a stationary shaft frame, at the ends of which an endless elastic belt has been mounted rotatably on bearings. Underneath the belt, at the roll nip, there is a shoe element that can be loaded. When the roll nip is closed, the endless belt is rotated on the stationary shaft frame and pressurised medium is conducted into the oil pocket between the belt and the shoe element to reduce friction. The pressurised medium is transferred into the oil pocket either by means of static or dynamic lubrication. In static lubrication, the medium passes through the shoe element into the oil pocket between the shoe element and the belt, and from there on via the leading edge and/or trailing edge of the shoe element back to the medium container. In dynamic medium circulation, the medium is transferred via the leading edge of the shoe element to the oil pocket and further via the trailing edge to the medium container. In both cases the medium cools the shoe element. The leading edge of the shoe element denotes the edge aligned with the longitudinal axis of the shoe element, over which the fibre web passing in the machine direction first passes in the roll nip, and the trailing edge of the shoe element, accordingly, denotes the edge aligned with the longitudinal axis of the shoe element over which the fibre web travelling in the machine direction passes as it leaves the roll nip.

In the shoe roll developed by the applicant, called a"symbelt roll", the loading means of the shoe element consist of hydraulic cylinders, each of which is formed of a piston/cylinder pair. Under the shoe element, there are two rows of hydraulic cylinders, the first one of which is located near the leading edge of the shoe element and the second near the trailing edge of the shoe element. The average nip pressure in the machine direction of the roll nip is controlled by loading of the shoe element by means of the subjacent hydraulic cylinders. Usually the oil pressure of the subjacent hydraulic cylinders loads the trailing edge of the shoe element more than the leading edge of the shoe element, so that the shoe element is inclined. The

inclination of the shoe element, i. e. the ratio of the load of the leading edge to that of the trailing edge, is called the tilt, i. e. the trailing edge is loaded more than the leading edge to a degree equalling the tilt. The tilt is hence a calculatory variable which is determined by the shoe element geometry and whose magnitude determines both the machine directional nip pressure distribution in a given roll nip and the maximum pressure prevailing in the roll nip.

In this application, the lower side of the shoe element implies the side of the shoe element on which the loading elements of the loading arrangement, such as the rows of hydraulic cylinders, for instance, are supported.

The oil pocket located on the upper surface of the shoe element is divided into sections by partitions, so that the lubricating oil conducted into these sections of the oil pocket do not admix substantially with the lubricating oil present in the adjacent sections.

The quality variables of the fibre web, i. e. the surface profile of the fibre web, such as the thickness, finish and smoothness of the paper web, can be acted on by means of several control variables in the shoe nip calender, of which only the main variables are mentioned below: - the tilt of the shoe element acts on the linear load profile of the roll nip transverse to the machine direction and on the linear load distribution in the machine direction, - the moisture and temperature of the fibre web can be controlled e. g. by blowing vapour onto the fibre web by means of steam boxes or by humidifying the fibre web by means of a water humidifier, - the temperature of the roll nip is controlled mainly by means of the temperature of the thermo roll and - the residence time of the fibre web in the roll nip is dependent on the path speed of the fibre web and on the length of the roll nip in the machine direction.

The control variables defined above enable control of the fibre web profiling in the roll nip of a shoe nip calender. However, when the aim is to achieve improved and more accurate controllability of the quality variables of the fibre web, the control

variables mentioned above are not adequate in all profiling conditions. In this case, a larger range of control variables would provide better means for more accurate control of the quality variables.

The invention is intended to eliminate the inconveniences of prior art.

Thus the chief purpose of the invention is to achieve a method for controlling the quality variables of a fibre web, in which new control variables can be used for controlling the quality variables.

The objectives above are attained by the method defined in claim 1, for instance.

In the method of the invention, the quality variables of the fibre web are controlled in a shoe nip calender comprising at least one shoe roll and a counter roll, between which there is a roll nip, and the tilt of the shoe element is used as a control variable in the control of the quality variables of the fibre web. The tilt of the shoe roll is determined at least partly on the basis of the temperatures measured in the shoe element of the shoe roll and/or the oil pressures measured in the oil pocket of the shoe element.

In a preferred embodiment of the invention, the quality variables of the fibre web are controlled by means of a multivariable controller, using an MPC control algorithm either in feedback control or feed-forward control. Feed-forward control is preferably performed by determining the values of one or more quality variables at given intervals as the fibre web speed changes from a first speed to a second speed over a period t. Said determined values of one or more quality variables are compared with the predicted set values of said one or more quality variables in order to obtain the difference value of one or more quality variables. After this, said difference values are used to determine new predicted set values for said quality variables, and said predicted quality variables are used to determine set values for one or more control variables.

The invention is based on the fact that the control of quality variables utilises the tilt i. e. inclination of the shoe element in a new manner; the local temperature and oil pressure differences of the shoe element are utilised in the determination of the tilt.

Local temperatures of the shoe element are determined at the leading edge and rear (trailing) edge of the shoe. element and the oil pressure is determined in different

sections of the oil pocket in the shoe element. These temperature and pressure measurement points are so densely distributed in the transverse direction to the machine direction, i. e. the cross direction (direction of the longitudinal axis of the shoe element) that the cross-directional temperature and (nip) pressure distribution of the shoe element is accurately determined. The temperature of the shoe element at the trailing edge increases more intensely than at the leading edge as the tilt is increased, and accordingly, the oil pressure in the oil pocket of the shoe element decreases as the tilt increases. These cross-directional oil pressure distribution data enable determination of the average machine-directional tilt and also of the tilt in each zone of the oil pockets of the shoe element.

When the calculated tilt values of the temperature and pressure distributions of the shoe element transverse to the machine direction, i. e. the cross directional values, are combined with the control of the oil pressure of the hydraulic cylinders loading the leading edge and trailing edge of the shoe element, new control variables are provided for control of the quality variables.

Among the additional advantages achieved with the invention, we may cite control of the tilt with higher precision than before, because the local temperatures of the shoe element and also pressures of the oil pocket sections are used for the control.

Measurement of the local temperatures of the shoe element and of the oil pocket sections also enables different tilts to be produced in the different zones of the shoe element, these tilts allowing modification of the linear load profile of the roll nip in the transverse direction of the roll nip, i. e. parallel to the longitudinal axis of the shoe element.

The invention is described in greater detail below with reference to the accompanying drawings.

Figure 1 is a schematic partly sectional view of the shoe element of a shoe nip calender and the hydraulic cylinders loading this.

Figure 2 shows the temperatures of the leading edge and trailing edge of the shoe element at different tilt values and at various lubricating oil temperatures.

Figure 3 shows the variation of the average temperature and pressure of the lubricating oil in the oil pockets of the shoe element in the transverse direction of the shoe element at different tilt values.

Figure 4 is a schematic view of the control method of the invention in a"feedback control arrangement".

Figure 5 is a schematic view of the control method of the invention in a"feed- forward control arrangement".

The following is a brief survey of the features of the invention illustrated by the figures.

Figure 1 illustrates the performance of the temperature and pressure measurements used for determining the inclination, i. e. tilt of the shoe element 4 of the shoe nip calender. The figure shows schematically in part section the shoe element 4 of the applicant's"sym-belt"shoe roll 2, which can be pressurised from below by means of the loading arrangement 5. Sensors for measuring the temperature of the shoe element and the oil pressure of the lubricating oil are provided both at the leading edge 4a and the trailing edge 4b of the oil pocket 4c in the shoe element.

Figure 2 shows the temperatures at the leading edge and trailing edge of the shoe element as a function of the loading ratio, i. e. tilt of the shoe element. T1-T4 are measurement points at the leading edge of the shoe element and T5-T8 are measurement points at the trailing edge of the shoe element.

Figure 3, in turn, illustrates the average temperatures in the zones parallel to the longitudinal axis of the shoe element and lubricating oil pressures as measured at different points of the shoe element when the load on the shoe element is altered (profiled) in the cross direction at various tilt values.

Figure 4 illustrates the control of quality variables 50 (thickness 51, polish 52, smoothness 53) of the fibre web performed as ordinary feedback control. This control manner comprises continuous measurement of quality variables and among the control variables 60, shoe element temperatures 61 affecting the value of the tilt 63 and oil pressures 62 in the oil pocket. The tilt is adjusted on the basis of the

difference values 50c of the quality variables and of said measurements of the local temperature and pressure of the shoe element acting on the tilt value.

Figure 5, in turn, shows the control of the quality variables of the fibre web when the speed V of the fibre web changes in"feed-forward"multivariable control using the control algorithm G (s), an"MPC"algorithm, in the process for minimising the cost function. The control manner comprises periodical measurements of the quality variables 50 of the fibre web and among the control variables, at least local temperatures 61 of each zone of the shoe element and the oil pressures 62 of the oil pocket sections, and these values are used for control of the process by changing the values of the quality variables and the control variables, such as the tilt.

The invention is described in greater detail below with reference to figures 1 to 5.

In the partly sectional shoe roll 2 shown in figure 1, two loading elements 5a ; 5a', 5a"of the loading arrangement 5 are supported on the lower surface of the shoe element 4, the loading elements in the type of shoe roll used being two rows of hydraulic cylinders, of which the figure shows the two first hydraulic cylinders of both the rows of hydraulic cylinders. The rows of hydraulic cylinders 5 extend fro the first end of the shoe element 4 to its other end. The first row of hydraulic cylinders 5a'is located under the leading edge 4a of the shoe element and the second row of hydraulic cylinders 5a"is located under the rear i. e. trailing edge 4b of the shoe element. These loading elements enable loading of the leading edge and the trailing edge of said shoe element under different pressures. The hydraulic oil is brought to the hydraulic cylinders in the rows of hydraulic cylinders in separate oil ducts 5b ; 5b', 5b"and each oil duct has a separate pressurising arrangement, such as an individual hydraulic motor or pump. The figure only illustrates the ducting leading to the first hydraulic cylinders of the oil ducts. There are also lubricating ducts passing through the shoe element (not shown in the figures) to the oil pocket 4c located on the upper surface of the shoe element. The oil pocket 4c is divided into sections 4d separated by partitions, the partitions preventing the lubricating oil from flowing between the oil pocket sections. Sensors 6b for measuring the pressure of the lubricating oil are provided in the oil pocket sections 4d. A plurality of sensors, marked 6a and 6b, respectively, for measuring the temperature of the zones are provided in the different zones of the leading edge 4a and the trailing edge 4b of the shoe element. The figure also shows part of the endless belt 7 sliding on the shoe element, with its direction of rotation indicated by an arrow with a full

head. The lubricating oil conducted into the oil pocket 4c forms a friction-reducing layer between the shoe element 4 and the endless belt 7 sliding on this when the roll nip is closed.

Figure 2 shows the temperature of the shoe element of figure 1 at different tilt values of the shoe. Because the leading edge and the trailing edge of the shoe element are at different angles to the endless belt sliding on the shoe element, the leading edge and the trailing edge will be heated in different ways, and the weight distribution in the machine direction (MD), i. e. the transverse direction to the shoe element, is not constant. The temperatures of the shoe element have been measured at the leading edge and the trailing edge of the shoe element. In the figure, the temperature curves T5-T8 denote temperatures of the trailing edge at lubricating oil temperatures 38,37, 36 and 35 Celsius with different shoe element tilts, and the temperature curves T1-T4, accordingly, the temperatures of the leading edge of the shoe element at the same lubricating oil temperatures and with the same shoe element tilts. The figure shows that the temperature at the trailing edge of the shoe element rises more abruptly than the temperature at the leading edge of the shoe element as the tilt, i. e. the inclination of the shoe element is increased. In this manner, it is possible to use the temperatures of the leading edge and the trailing edge of the shoe element and the mutual development of these temperatures in the determination of the average inclination of the shoe element in the machine direction.

Figure 3 shows the oil pressure of the lubricating oil of the shoe element in figure 1 in the oil pocket sections and the average temperature of the shoe element in the cross-direction of the shoe element. The curves P1, P2, P3 and P4 describe the oil pressures in the oil pocket sections of the shoe element with a tilt of 1.05, 1.15, 1.25 and 1. 35, respectively. The curves Tl, T2, T3 and T4 describe the average temperatures of the shoe element with tilt values 1.05, 1.15, 1.25 and 1.35, respectively. The curves show that the lubricating oil pressure in the oil pocket sections drops as the inclination of the shoe element, i. e. the tilt increases. Thus the inclinations in each zone of the shoe element can be determined on the basis of oil pressure values measured at different points of the oil pocket and of temperatures measured at different points of the leading edge and the trailing edge.

Now the temperature curves of the leading edge and the trailing edge of the shoe element 4 in figure 1 and the oil pressure curves of the oil pocket sections and the

temperature differences between the leading edge and the trailing edge of the shoe element can be utilised for accurate control of the tilt value. Because each hydraulic cylinder located under the shoe element can additionally be pressurised under different pressures, individual load can be provided in each zone of the shoe element. Thus, for instance, the hydraulic cylinders located under the trailing edge 4b or leading edge 4a of the shoe element, can, if desired, be used to load the leading edge and the trailing edge of the shoe element in the different zones in slightly different ways in the cross direction. Thus, different loading of the leading edge and the trailing edge located in different zones of the shoe element may serve to change the nip pressure profile prevailing in a given roll nip in the cross direction and thus the profiling pressure of the fibre web W passing between the endless belt 7 sliding on the shoe element and the counter roll. Hence temperature and oil pressure measurements of the shoe element 4 enable accurate control of the linear load profile of the roll nip between the shoe roll 2 and the heated counter roll to the desired value, thus increasing the control potential of fibre web profiling.

The temperature and oil pressure curves of the shoe element described above and the temperature differences between the leading edge and the trailing edge of the shoe element always allow also calculation of the average temperature of the shoe element 4 prevailing in the machine direction of the roll nip and the average oil pressure in the oil pocket, thus allowing for accurate tilt value determination in the machine direction.

If the shoe nip calender 1 comprises a plurality of roll nips, the tilt value of each roll nip is controlled separately by observing as above the oil pressure of the lubricating oil in the oil pocket 4c in each shoe element 4 of the shoe roll 2 and the temperatures of the leading edge and the trailing edge 4a, 4b of the shoe element.

Measurement as above of the cross-directional temperatures of the shoe element 4 and the lubricating oil pressures in different sections 4d of the oil pocket 4c enables control of the magnitude of the nip pressure distribution of the roll nip between the shoe roll and its counter roll in the machine direction and the shape of the cross- directional linear pressure profile. If the shoe element 4 and/or the endless belt 7 has a specific maximum temperature not to be exceeded for reasons of material technology or specific paper grades, for instance, these measurements provide for accurate control that this maximum temperature is not exceeded in any part of the shoe element and/or the belt.

Below is a more detailed description of how to utilise temperature and pressure measurements of the shoe element 4 as an active online measurement variable of the shoe nip calender 1 both in feed-forward and feed-back controls for controlling the maximum temperature of the roll nip, the shape of the linear pressure profile in the cross-direction and the linear pressure distribution and the nip pressure in the machine direction.

Figure 4 shows the arrangement of the invention for controlling the nip load on a roll nip. The figure includes a schematic drawing of a shoe nip calender 1 comprising one roll nip N between the shoe roll 2 and the heated counter roll 3. The fibre web W passes in the roll nip in the direction indicated by the arrow from the left to the right. After the roll nip, one or more of the quality variables 50 (here thickness 51, smoothness 52, polish 53) are measured continuously or at given intervals. The measurements can be performed at a constant point of the fibre web or at a sensor passing on (traversing) the fibre web. Part or all of the control variables 60 are also periodically measured for adjustment of the control. The tilt 63 of the shoe element is used as control variables 60 in the process, either defined as average tilt in the machine direction between the leading edge and the trailing edge of the shoe element, or as tilt in each zone of the shoe element in the cross direction.

For determination of the tilt values, the control variables used consist of measurements 61 of the temperature at the leading edge and the trailing edge of the shoe element, as shown in figure 1, performed at given points in the cross direction of the shoe element, and oil pressure measurements 62 of the oil pocket of the shoe element. The oil pressure and temperature measurements mentioned above enable calculation of the average nip load of the roll nip as well. The nip pressure 64 generated by the nip load may be either observed separately or determined by the temperature and pressure measurements mentioned above. Other conceivable control variables to be measured include the moisture and temperature of the fibre web.

The measured values 50 ; 50a, 5 la ; 52a, 53a of the quality variables of the fibre web are compared with the predetermined set values 50b ; 51b, 52b, 53b of the same quality variables, yielding the difference values 50c ; 51c, 52c, 53c of said quality variables. Depending on the control algorithm G (S) (unit controller, which usually is a"PID control"or a multivariable controller), one or more control variables are used for control of the quality variable values. The objective in PID control is to

control the quality variables 50 of the fibre web to the desired values by means of changes of one control variable 60, in other words, the aim is to minimise the difference values 50c of the quality variables. Multivariable control aims at minimising the difference values 50c of the quality variables by means of simultaneous control of a plurality of control variables 60. Before the control variable is changed, the present value of the control variable must be known, and it can be determined in the process, for instance. Regardless of the control method, the effect of a given control variable 60 on the quality variable/quality variables 50 must be known, via a response model, as a function, table or curve, for instance. In the case of a multivariable controller, the mutual cross-effect of the control variables must usually be known as well. Subsequently, the control variables 60 used are changed on the basis of the selected control algorithm G (s). A preferred control strategy uses a multivariable controller, e. g. the MPC controller described in detail in example 5 below, allowing minimum and maximum limits to be set for the control variables 60 of tilt 63 of the shoe element and/or the temperature 61 of the leading edge and the trailing edge of the shoe element and/or the pressure 62 of the lubricating oil. A multivariable controller allows simultaneously for the effect of a plurality of control variables 60 (linear load, temperature of the thermo roll, tilt and vapour amount) on the quality variables 50 of the fibre web, such as a paper web (thickness/polish/smoothness).

In the feedback control illustrated in figure 4, the paper quality variables 50 are continuously measured. One of the control variables 60 used is the inclination, i. e. tilt 63 of the shoe element. Among the control variables acting on the tilt, the measurements include the lubricant oil pressure 62 in the oil pocket of the shoe element at several points of the oil pocket of the shoe element and of the temperature of the shoe element at several points of the leading edge and trailing edge of the shoe element.

The conceivably optimal values 61,62, 63 for the control variable 60 are calculated on the basis of the response models in order to obtain the desired nip pressure distribution in the machine direction, nip pressure profile in the cross direction and the condition of the shoe element (temperature/pressure). The difference values 50c between the values 50a of the quality variables measured on the fibre web and the set values 50b are used as a basis for changing the average or zone-related tilt 63 of the shoe element as desired, by adjusting the tilt of the shoe element by means of variation of the pressure of the loading elements under the leading edge and the

trailing edge of the shoe element, such as hydraulic cylinders. The pressure of the hydraulic cylinders is used for controlling also the nip load and hence the nip pressure as well. Additionally, measurements of the temperature and oil pressure of the shoe element can be used as a basis for changing the flow of lubricating oil entering the oil pocket of the shoe element and/or the temperature of the lubricating oil. The maximum temperature of the shoe element is then determined by the temperature of the lubricating oil and the tilt of the shoe element.

In the feed-forward control shown in figure 5, quality variables 50 are controlled as the machine speed V changes, i. e. as the fibre web speed accelerates or decelerates substantially. The control algorithm in the figure is an MPC controller, which is a multivariable controller allowing control windows (min. /max. values) to be set for the control variables for the duration of the control. An MPC controller minimises the square-law cost function of the difference between the set values of the quality variables (predicted or constant set values) and the values measured for the same quality variables in the process, and it also allows for compensation of dead process time. The control algorithms of MPC controllers, including the control algorithms of multivariable controllers, have been described in detail e. g. in the publication Aiche Symposium, vol. 93-97, pp. 232 to 256, California 1991.

Figure 5 shows a shoe nip calender 1 comprising two roll nips N; N' ; N", a fibre web being profiled in the roll nips N', N"of the shoe element. A thermo roll 3; 3', 3"is provided opposite the shoe roll 2,2', 2"in the roll nips. As the speed V of the fibre web changes from a first speed to a second speed, the process is controlled by feed-forward control. In this control, the quality variables 50 of the fibre web are measured at suitable intervals either point wise at one location of the fibre web W or by means of a traversing sensor. Since the controller G (s) has anticipatory control, which is intended to compensate for rapid changes, all the quality variables 50 need not necessarily be measured each time, but can be anticipated by approximative models. These values 50a measured for the quality variables 50 are compared with the predicted values 50b'and 50b"of the same quality variables, and a new predicted set value of the quality variable is calculated on the basis of this. When the predicted set values of the quality variables are the same as their constant set values 50b, the control is stopped. In this control, feedback is thus performed by comparing the measured values 50a of the quality variable/quality variables with the predicted set values 50b'and 50b"of the same quality variables. The effect of the tilt of the shoe element, the temperatures of the leading edge and the trailing edge of

the shoe element and the pressure of the lubricant oil in the oil pocket of the shoe element on each quality variable 50 is known either via a response model, as an approximative function, as a table or a curve. Using these action models the tilt of the shoe and the machine-directional load on the roll nip are controlled by means of the hydraulic cylinders located under the leading edge and the trailing edge of the shoe in such a way that each predicted value of a quality variable 50 of the fibre web corresponds to a given set value of the machine-directional nip pressure of the tilt and/or roll nip, which is used as control variable 60. The control variables mentioned above can be relatively rapidly made by means of hydraulic cylinders; within 5 to 10 seconds, the set values can be changed either by controlling directly the control circuits loading the hydraulic actuators, such as hydraulic pumps and/or hydraulic motors and valves, which are located in the flow ducts and are usually hydraulically operated. In addition, the flow rate and temperature of the lubricating oil in the oil pocket can be changed.

Optionally, the change set values of the control variables 60 can be performed as an automatic staggered change. In this case, the set values mentioned above can be changed in a linear manner by estimating the effect of the speed change of the fibre web on the quality variables and by adjusting the control model precision only periodically. An anticipatory control strategy eliminates quality variations caused by changes in the machine speed.

Only a few embodiments of the method of the invention have been described above, and it is obvious to those skilled in the art that the invention can be implemented in many other ways without departing from the scope of protection of the invention as defined in the claims.

Thus, the shoe roll may have a design in which its shoe element is pressurised with one single row of hydraulic cylinders, and then the value of the tilt of the shoe element cannot be controlled in on-line control, because the tilt value is constant.

The quality variables of a fibre web of a shoe nip calender having such a shoe roll are controlled by means of the tilt in the same way as the control of a shoe nip calender having a shoe roll equipped with two rows of cylinders as described in the examples above, but the tilt of the shoe element is controlled only when the fibre web is not being calendered in the roll nips.

The number of oil pockets on the upper surface of the shoe element and of sections in the oil pockets may vary considerably. Yet one should note that even though the oil pocket would comprise one single oil pocket section (with no partitions 4d in the oil pocket), the pressure differences at different points of the oil pocket can still be observed by measurements.