Data processing: generic control systems or specific application – Generic control system – apparatus or process – Optimization or adaptive control
Reexamination Certificate
2000-10-02
2003-06-24
Follansbee, John (Department: 2121)
Data processing: generic control systems or specific application
Generic control system, apparatus or process
Optimization or adaptive control
C700S047000, C700S030000, C700S031000, C700S126000, C706S023000, C706S904000
Reexamination Certificate
active
06584366
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a method and arrangement for preparing a neural model of a paper winding device, such as a drum cutter.
2. Description of the Related Art
Given the production of paper, it has proven successful to store the paper in the form of drums, which are also referred to as tambour or coil, since very large paper quantities thereby can be stored crease-free in the smallest space for bearing, transporting and processing.
A paper winding device contained in the drum cutter has the task to unwind a broad coil of paper and to develop paper webs corresponding to the requests of the customers, whereby conventional paper widths can certainly amount to 9.5 m. For this purpose, these are cut in longitudinal direction and the narrower paper webs arising thereby are wound up again. Doubling roll coilers are frequently utilized due to their technical preferences for the winding-up (one is schematically shown in FIG.
1
). This winding process critically influences the quality of the paper on the wound rolls, this quality being strongly dependent on the tensions in the paper occurring in the rolls. The tangential and radial tensions arising in the paper coil, among other things, are influenced by the following influencing and control quantities of the paper winding device and the paper:
by their moments and speed given the center drive;
by its moment and speed given the doubling roll drive;
by the winding radius, the line force in the nip, adjustments of the friction dampers and the width of the coils given the winding stations;
by geometrical influencing quantities, such as the doubling roll radius, the width of the tambour, and the radius of the sleeve onto which the paper coils are wound;
by the properties of the paper, such as the modulus of elasticity, the G.M.S per density, the roughness, the smoothness, the moisture, the porosity, as well as by the elongation at rupture;
by other influencing quantities, such as the web strength, the roughness of the doubling roll, the friction value of the doubling roll for each paper type, as well as by the modulus of elasticity of the sleeve onto which it is wound.
Since these influencing and controlling quantities are also dependent on environmental conditions, such as the humidity and the temperature, such winding processes are very difficult to describe analytically. For these reasons, it is particularly difficult to meet the quality requirements with respect to such coils and to constantly keep them. Expert knowledge is still utilized thereby with respect to the prior art, in particular Predominantly, it is to be assured that reproducible coils having an optimal coil hardness structure arise during the winding process. Furthermore, a lateral progressing, which is also referred to as telescoping of the roll, is to be prevented. The winding process is to be particularly rid off tears and broken picks, as well as plastic deformations.
The controlling quantities for such paper winding devices and particularly for drum cutters have been hitherto fixed, in the form of set value tragjectories in an initial operation phase by initial operation engineers, manually and on the basis of their expert knowledge. This requires great technological knowhow, skills and time. Generally, a separate set of set value trajectories of the paper winding devices is fixed for each paper type of the product spectrum and is stored in the, for example, process computer, which controls the drum cutter. In a later operation, the set value trajectories are read out of the memory and are utilized for controlling the winding device. A specific quality level is achieved as a result of this course of action; however, the quality that can be obtained thereby is limited by the following factors:
it cannot be assured that optimal control trajectories are always utilized, since there is no mathematical model for determining optimal control trajectories. On a case by case basis, sub-optimal control trajectories therefore can be applied.
drifts of measuring devices or as a result of a tear remain unconsidered.
variations of the paper parameters within one type, such as the G.S.M or the elasticity module, remain unconsidered.
When the winding quality worsens below an acceptable value as a result of the cited effects during the operation of a drum cutter or of another paper winding device, the driver of the device must manually intervene for correction purposes. General theoretical basics for the processes occurring during the paper winding are described in H.-J. Schaffrath, F. Hibinger and L. Goettsching “Numerische Berechnung von Spannungsverlaeufen und Wickelhaerte in einer Papierrolle”, pages 350-361, No. 6, 1994. Methods for the automated preparation of control parameters for paper winding devices and particularly for drum cutters are not yet known.
SUMMARY OF THE INVENTION
The present invention is based on the object of providing a method and a device for the neural modeling of paper winding devices and particularly for the optimization of the control parameters of these devices.
This and other objects and advantages of the invention are provided by a method for the neural modeling of a paper winding device, wherein influencing and control quantities of a paper winding device are provided in a first step for the training of a neural network, in that these are measured and are stored at least depending on the wound number of layers and the appertaining coil radius of the paper coil, wherein the web strength in the paper web or a quantity correlated with it, depending on the influencing and control quantities of the paper winding device, is determined as target quantity for the neural network from the relationship between the measuring quantities coil radius and appertaining number of layers of the paper coil, and wherein the neural network is trained as model of the paper winding device via a conventional learning method at least with influencing and control quantities of the paper winding device as input quantities and the web strength dependent thereon or, respectively, the quantity correlated with the web strength as output quantities.
In the method, the influencing and control quantities are measured at a real paper winding device and/or at the paper, and/or the environment of the paper winding device or of the paper. The influencing and control quantities are measured dependent on time.
In one embodiment, the method for generating optimal control quantities for a paper winding device depending on a desired winding quality of the paper coil, wherein influencing quantities and control quantities of the paper winding device to be optimized are supplied in a first optimization step to the neural model prepared according to one of the claims 1 through 3 and wherein a first web strength or, respectively, a first quantity correlated with the first web strength is determined therefrom with the aid of the model, wherein a first winding quality is determined in a second step from the known relationship between the web strength or, respectively, the quantity correlated with the web strength and winding quality from the first web strength or, respectively, the correlated first quantity, and wherein the first winding quality is compared to the desired winding quality and wherein at least second control quantities of the paper winding device to be optimized are supplied in a further optimization step to the model, whereby this course of action is kept until the winding quality determined from the two steps above sufficiently accurately corresponds to the desired winding quality.
The change of the winding quality is determined depending on the change of the control quantities of two adjacent optimization steps and wherein it is concluded from this change in which way the control quantities are to be changed for the next optimization step in order to arrive at the desired winding quality. The control trajectories may be generated as a time-dependent sequence of control variables.
An operating method for a paper windi
Liepold Helmut
Schäffner Clemens
Wienholt Willfried
Baker & Botts LLP
Barnes Crystal J
Follansbee John
Siemens Aktiengesellschaft
LandOfFree
Method and arrangement for neural modeling of a paper... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and arrangement for neural modeling of a paper..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and arrangement for neural modeling of a paper... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3092137