Method to restore the functional state of an automatic machine for the production or the packaging of consumer products

Patent metadata

Field	Value
Publication number	US-12130615-B2
Application number	US-202017610138-A
Country	US
Kind code	B2
Filing date	May 29, 2020
Priority date	May 30, 2019
Publication date	Oct 29, 2024
Grant date	Oct 29, 2024

Abstract

A method to restore the functional state of an automatic machine for the production or the packaging of consumer products. The method comprises the steps of: storing a first knowledge base containing a plurality of problematic operating conditions having a corresponding known fault; training the data processing system by associating ate least one solution to each problematic operating condition and storing a second knowledge base containing the solution; identifying a current operating condition; searching, among all the problematic operating conditions stored in the first knowledge base, for the problematic operating condition that is the most similar to the current operating condition; and communicating to an operator the solution associated with the corresponding problematic operating condition and stored inside the second knowledge base.

First claim

The invention claimed is:

1. A method to restore the functional state of at least part ( 3 ) of an automatic machine ( 1 ) for the production or the packaging of consumer products;
the automatic machine ( 1 ) comprises a plurality of movable operating members, a control unit ( 4 ), a data processing system ( 5 ), a writable memory ( 6 ) connected to the data processing system ( 5 ), and a user interface device ( 7 ) connected to the data processing system ( 5 );
the method comprises the steps of:
identifying, for one time only, a list of all the operating variables ( 8 ) characterizing the operation of the automatic machine ( 1 );
storing, for one time and inside the writable memory ( 6 ), a first knowledge base ( 9 ) containing a plurality of problematic operating conditions ( 10 ), each consisting of a set of values associated with the operating variables ( 8 ) and of at least one corresponding known fault ( 11 );
storing, for one time and inside the writable memory ( 6 ), a second knowledge base ( 13 ) containing a corresponding solution ( 12 ) for each known fault ( 11 ) of the first knowledge base ( 9 );
identifying, through the control unit ( 4 ) and following or prior to an unknown fault ( 15 ) of the automatic machine ( 1 ), a current operating condition ( 14 ) consisting of a set of current values of the operating variables ( 8 );
searching, among all the problematic operating conditions ( 10 ) stored in the first knowledge base ( 9 ), for the problematic operating condition ( 10 ) that is the most similar to the current operating condition ( 14 ); and
communicating, in case inside the first knowledge base ( 9 ) there is at least one problematic operating condition ( 10 ) matching the current operating condition ( 14 ), to an operator (O) the solution ( 12 ) associated with the matching problematic operating condition ( 10 ) and stored inside the second knowledge base ( 13 ); wherein the data processing system ( 5 ) asks the operator (O) for a confirmation of the effectiveness of the communicated solution ( 12 ); if the communicated solution ( 12 ) is not decisive and there are no alternative solutions ( 12 ) for the unknown fault ( 15 ) inside the second knowledge base ( 13 ), the data processing system ( 5 ) actively interacts with the operator (O), through the user interface device ( 7 ), obtaining additional information and processing one or more additional alternative solutions ( 12 ); the method also comprises a machine learning step ( 30 ) performed by the data processing system ( 5 ) on the basis of the interaction between the data processing system ( 5 ) and the operator (O), performed through the user interface device ( 7 ).

2. The method according to claim 1 , wherein, if the communicated solution ( 12 ) is not decisive, the data processing system ( 5 ) communicates to the operator (O) at least one alternative solution ( 12 ) for the unknown fault ( 15 ), if it exists inside the second knowledge base ( 13 ).

3. The method according to claim 1 , wherein the data processing system ( 5 ) asks the operator (O) for confirmation of the effectiveness of said one or more additional alternative solutions ( 12 ).

4. The method according to claim 1 , wherein, during the interaction, the data processing system ( 5 ) asks the operator (O) structured questions ( 17 ).

5. The method according to claim 1 , wherein, during the interaction, the operator (O) indicates to the data processing system ( 5 ), through the user interface device ( 7 ), the part ( 3 ) of the automatic machine ( 1 ) suffering from the unknown fault ( 15 ) based on a three-dimensional model ( 16 ) of the automatic machine ( 1 ).

6. The method according to claim 1 , wherein the data processing system ( 5 ) processes the additional alternative solutions ( 12 ) based on artificial intelligence algorithms, communicating to the operator (O) at least one solution ( 12 ) to a known fault ( 11 ) associated with a different problematic operating condition ( 10 ), which is, however, similar to the current operating condition ( 14 ).

7. The method according to claim 6 , wherein the data processing system ( 5 ) updates the first and/or the second knowledge base ( 9 , 13 ) and/or establishes connections between the elements inside the first and the second knowledge base ( 9 , 13 ) based on the additional alternative solutions ( 12 ) processed following the interaction with the operator (O).

8. The method according to claim 1 , wherein the machine learning step ( 30 ) takes place by means of a direct connection between the user interface device ( 7 ) and the data processing system ( 5 ).

9. The method according to claim 1 , wherein the machine learning step ( 30 ) takes place by means of a remote connection, in particular thorough distributed architecture, between the user interface device ( 7 ) and the data processing system ( 5 ).

10. The method according to claim 1 , wherein the first and the second knowledge base ( 9 , 13 ) are shared by different automatic machines ( 1 ), in particular of the same type and/or designed to produce the same product.

11. The method according to claim 1 , wherein, if none of the alternative solutions ( 12 ) processed by the data processing system ( 5 ) solve the unknown fault ( 15 ), the data processing system ( 5 ) learns, from the answers ( 18 ) of the operator (O) to the structured questions ( 17 ), a new solution ( 20 ) for the unknown fault ( 15 ) and stores this new solution ( 20 ) inside the second knowledge base ( 13 ), associating it with the current operating condition ( 14 ), which is stored inside the first knowledge base ( 9 ) as a new problematic operating condition ( 10 ).

12. An automatic machine ( 1 ) for the production or the packaging of consumer products; the automatic machine ( 1 ) comprising:
a plurality of movable operating members, each capable of assuming a plurality of different positions;
a control unit ( 4 );
a writable memory ( 6 ); and
a data processing system ( 5 );
the automatic machine ( 1 ) is characterized in that a first and a second knowledge base ( 9 , 13 ) containing a plurality of possible problematic operating conditions ( 10 ) of the automatic machine ( 1 ) and at least one solution ( 12 ) for each problematic operating condition ( 10 ), respectively, are stored inside the memory ( 6 ); the automatic machine ( 1 ) being designed to carry out the method according to claim 1 .

Assignees

• Gd Spa

Inventors

• Albertini Daniela
• Bertuzzi Gianluca
• Federici Luca

Classifications

• G05B2219/23253

  Expert system · CPC title

• G05B19/0426 Primary

  Programming the control sequence · CPC title

• G05B23/0243

  model based detection method, e.g. first-principles knowledge model · CPC title

• G05B23/0275

  Fault isolation and identification, e.g. classify fault; estimate cause or root of failure · CPC title

• G05B23/0229 Primary

  knowledge based, e.g. expert systems; genetic algorithms · CPC title