Method and device for determining relevant variables in the processing of textile structures

• US 6,263,257 B1
• Filed: 08/22/1995
• Issued: 07/17/2001
• Est. Priority Date: 08/22/1994
• Status: Expired due to Fees

• Current Assignee: Zellweger Luwa AG (Hesta AG)
• Original Assignee: Zellweger Luwa AG (Hesta AG)
• Inventors: Aemmer, Peter F.
• Primary Examiner(s): Cuchlinski, Jr., William A.
• Assistant Examiner(s): MARC, MCDIEUNEL
• Application Number:
  US08/517,628
• Time in Patent Office: 2,156 Days
• Field of Search: 700/48, 700/31, 700/150, 700/37, 700/23, 700/47, 700/44, 700/139, 700/130, 700/140, 700/143, 700/144, 706/23, 706/25, 706/903, 706/906, 706/44
• US Class Current: 700/139
• CPC Class Codes: G05B 13/027 using neural networks only
• CPC Class Codes: G05B 17/02 electric
• CPC Class Codes: G05B 19/41875 characterised by quality surveillance of production
• CPC Class Codes: G05B 2219/32193 Ann, neural base quality management
• CPC Class Codes: G05B 2219/32338 Use new conditions for model, check, calculate if model meets objectives
• CPC Class Codes: G05B 2219/32385 What is simulated, manufacturing process and compare results with real process
• CPC Class Codes: G05B 2219/33036 Error back propagation
• CPC Class Codes: G05B 2219/45191 Spinning, web spinning
• CPC Class Codes: Y02P 90/02 Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS]

Abstract

A device for the simulation of spinning machines with a view to their optimum economic use and operation. Starting from material properties of a preproduct (1) determined by measurement, as well as from configuration and setting parameters (3) of such a spinning machine, material properties (2) of the intermediate or output products which are being produced, which properties can be determined by measurement, are predicted using a process model describing the behavior of the spinning machine, by a simulation device. The process model on which a method being based is presented by a neural network. The coefficients determining the actual behavior of this neural network area calculated from a set of sample data in a training phase. This sample data and/or from otherwise conclusively predicted properties of the behavior of such a spinning machine.

Claims

1. A method for simulating the processing of input products in a spinning machine by means of a process model of nonlinear equations that is embodied in a neural network, comprising the steps of training the neural network by inputting data relating to (i) said input products, (ii) output products produced by said processing, and (iii) configuration parameters of the processing operation carried out in the spinning machine, and operating the trained neural network to produce estimates of the operation of the spinning machine.

2. The method of claim 1 wherein during said operating step first and second groups of data are input to the neural network and a third group of data is determined therefrom, wherein said three groups of data are respectively formed by one of (a) data relating to input products, (b) data relating to output products, and (c) data relating to configuration parameters of the spinning machine.

3. The method of claim 2 further including the step of repeatedly modifying data from at least one of said first and second groups to determine said third group of data.

4. The method of claim 2 wherein said third group of data comprises data relating to configuration parameters of the spinning machine, and further including the step of setting parameters of the spinning machine in accordance with said third group of data.

5. The method of claim 1 wherein said input data includes data determined by experimental measurements and data derived from said measurement data.

6. The method of claim 5 wherein the input data is contained within a partial range of an entire range of measurement data.

7. The method of claim 1 further including the step of identifying specific data relating to input products and output products as being associated with particular input products and output products, respectively, and associating specific configuration parameters with particular output products processed by the spinning machine.

8. The method of claim 7 wherein the data is obtained by measuring the input and output products with measuring devices that are independent of and physically separate from the spinning machine.

9. The method of claim 1 further comprising the steps of forming first and second process models that respectively pertain to first and second spinning machines, and providing output data from said first process model as input data to said second process model.

10. A device for simulating the processing of products in a spinning machine, comprising:

means associated with the spinning machine for producing data relating to the operation of the spinning machine;

a computer having a data input/output unit;

means for transmitting data between said data producing devices and said computer;

a program memory in said computer storing data pertaining to a model structure that is embodied in a neural network; and

a database containing data relating input products processed by the spinning machine, output products produced by the spinning machine, and configuration parameters for the spinning machine, for input to said neural network.

11. The device of claim 10 further including measuring devices that are physically separate from said spinning machine for generating data that is contained in said database.

12. The device of claim 11 further including a second computer and means for transmitting data from said measuring devices to said second computer, and wherein each of said transmitting means comprises a field bus, and said field busses are connected to one another.

13. A method for simulating the operation of a spinning machine, comprising the steps of:

training a neural network with input data relating to (i) input products that are processed by spinning machines, (ii) operating parameters of spinning machines, and (iii) output products produced by spinning machines, to thereby define coefficients for a simulation model; and

subsequently operating the training neural network using at least two of said three types of data (i), (ii) and (iii) as input data and said model coefficients to produce output data which describes the simulated operation of a spinning machine.

14. The method of claim 13 wherein said output data comprises the third type of data (i), (ii) or (iii).

15. The method of claim 14 wherein said input data that is used in the operation of the trained neural network comprises data relating to (i) input products that are processed by spinning machines, and (ii) operating parameters of spinning machines, and said output data comprises data relating to (iii) output products produced by spinning machines.

16. The method of claim 14 wherein said input data that is used in the operation of the trained neural network comprises data relating to (i) input products that are processed by spinning machines, (iii) output products produced by spinning machines, and said output data comprises data relating to (ii) operating parameters of spinning machines.

17. The method of claim 13 comprising the steps of training a neural network with one set of input data to define coefficients for a first simulation model, training a neural network with a different set of input data to define coefficients for a second simulation model, operating a neural network in accordance with the coefficients for the first simulation model to produce output data, and operating a neural network in accordance with the coefficients for the second model, using at least some of the output data produced by the first simulation model as input data, to produce further output data.