Method for controlling process events using neural network

• US 6,314,413 B1
• Filed: 08/13/1998
• Issued: 11/06/2001
• Est. Priority Date: 08/13/1997
• Status: Expired due to Term

• Current Assignee: Ralf Otte
• Original Assignee: ABB Patent GmbH (ABB Limited)
• Inventors: Otte, Ralf
• Primary Examiner(s): Davis, George B.
• Application Number:
  US09/133,931
• Time in Patent Office: 1,181 Days
• Field of Search: 706/15, 706/23, 706/907
• US Class Current: 706/15
• CPC Class Codes: G05B 13/027 using neural networks only
• CPC Class Codes: G06N 3/04 Architecture, e.g. interconnection topology
• CPC Class Codes: Y10S 706/907 Power plant

Abstract

The invention relates to a method for controlling process events of a technical plant. In order to permit a simultaneous and coherent assessment of relevant process variables of the plant, it is proposed to use a neural analysis on the basis of self-organizing neural maps to evaluate the relevant process variables in relation to one another by realizing a topology-maintaining nonlinear projection of data from the relevant process variables onto a multidimensional neural map.

Claims

1. A method for controlling process events of a technical plant, which comprises:

collecting variables relevant to a process;

realizing a topology-maintaining nonlinear projection of data from the variables onto a multidimensional self-organizing neural map;

displaying a trajectory of process events by connecting visualized process states graphically;

displaying reference channels with adjustable tolerance widths on the map;

evaluating the variables in relation to one another using the neural map;

controlling process events in accordance with the evaluating step;

evaluating sequences of process states using the reference channels;

detecting if a trajectory deviates from one of the reference channels;

evaluating a winner rate neuron representing a state; and

bringing the process back into the reference channel.

2. A method of controlling process events of a technical plant, comprising:

selecting process variables;

developing a self-organizing, neural network in a state space of a plant, on a basis of a self-organizing neural algorithm using process values of the process variables;

representing the self-organizing neural network as a neural map;

determining plant states from the process variables;

producing a reference channel of permissible process events by projecting the plant states onto the neural map;

recording the process events by linking the neural map to a process of the plant;

displaying a trajectory on the neural map;

monitoring whether the trajectory remains within or deviates from the reference channel; and

steering the process of the plant over the neural map in real terms in the state space.

3. The method of controlling plant states according to claim 2, wherein the projection step comprises projecting the plant states onto the neural map with a winner-takes-all algorithm.

4. The method of controlling plant states according to claim 2, wherein the projection step comprises projecting the plant states onto the neural map with a winner-takes-most algorithm.

5. The method of controlling plant states according to claim 2, wherein the steering step comprises steering the process over the map by using a process steering means.

6. The method of controlling plant states according to claim 2, which further comprises conditioning the process values of the process variables prior to the developing step.