Method for determining spraying parameters for a paint spraying unit

• US 6,507,803 B1
• Filed: 07/31/2000
• Issued: 01/14/2003
• Est. Priority Date: 07/31/1999
• Status: Expired due to Term

• Current Assignee: ABB Research Ltd. (ABB Limited)
• Original Assignee: ABB Research Ltd. (ABB Limited)
• Inventors: Eickmeyer, Dietmar, Börner, Gunter
• Primary Examiner(s): Hilten, John S.
• Assistant Examiner(s): PRETLOW, DEMETRIUS R
• Application Number:
  US09/630,149
• Time in Patent Office: 897 Days
• Field of Search: 702/170, 702/127, 706/15, 706/16, 706/25, 118/50.1, 118/300, 118/663-677, 118/712, 118/629, 361/227, 361/225, 427/7, 427/9, 427/10, 427/8, 427/446-456, 239/63, 239/708
• US Class Current: 702/170
• CPC Class Codes: B05B 12/00 Arrangements for controlling delivery; Arrangements for controlling the spray area
• CPC Class Codes: G05B 13/027 using neural networks only

Abstract

A method for determining spraying parameters that are suitable as input values for a paint spraying unit that can electrostatically charge a liquid paint. In this case, at least one artificial neural network is used to determine the spraying parameters, an output of such a neural network being available for each spraying parameter. A suitable number of real measured values are fed to the one neural network or a plurality of neural networks as input values, initially in a learning phase. The measured values further contain associated real spraying parameters in addition to a paint thickness distribution in the form of discrete values. Input values are fed to the one neural network or a plurality of neural networks in the application phase. The input values being the result of an analysis of the paint thickness distribution of a targeted, that is to say prescribed, spraying result.

Claims

1. A method for determining spraying parameters suitable as input values for a paint spraying unit that can electrostatically charge liquid paint, which comprises the steps of:

using at least one artificial neural network to determine the spraying parameters, and the artificial neural network having an output available for each of the spraying parameters;

feeding a number of real measured values to the at least one artificial neural network as input values, initially in a learning phase, the real measured values containing associated real spraying parameters in addition to a paint thickness distribution in a form of discrete values; and

feeding the input values to the at least one artificial neural network in an application phase, the input values being a result of an analysis of the paint thickness distribution of a prescribed spraying result.

2. The method according to claim 1, which comprises determining the real measured values, to be input in the learning phase, at least partly by a mathematical model of the paint spraying unit.

3. The method according to claim 1, which comprises that for each of the spraying parameters to be determined, using one artificial neural network having an output value bearing a fixed relationship to the spraying parameter to be determined.

4. The method according to claim 1, wherein the at least one neural network has a plurality of outputs which respectively correspond to one of the spraying parameters to be determined.

5. The method according to claim 1, which comprises using a multilayer perceptron trained with an aid of a backpropagation method as the at least one artificial neural network.