Piping diagnostic device, piping diagnostic method, discriminant-problem optimizing device, discriminant-problem optimizing method, criteria distribution generating device, criteria distribution generating method, and storage medium

Patent metadata

Field	Value
Publication number	US-11614205-B2
Application number	US-201716078133-A
Country	US
Kind code	B2
Filing date	Feb 16, 2017
Priority date	Feb 26, 2016
Publication date	Mar 28, 2023
Grant date	Mar 28, 2023

Abstract

A piping diagnostic device according to an exemplary aspect of the present invention includes: criteria distribution generating means for generating a criteria distribution that is statistical data of criteria data of piping based on construction information about the piping, design information about the piping, and material information about the piping; changed-state distribution generating means for generating a changed-state distribution that is statistical data of changed-state data of the piping that have changed due to aging based on at least either vibration or dynamic pressure of the piping that have changed due to aging; measurement means for measuring at least either vibration or dynamic pressure of the piping; and determining means for determining deterioration of the piping based on the criteria distribution, the changed-state distribution, and at least either vibration or dynamic pressure of the piping.

First claim

What is claimed is:

1. A piping diagnostic device comprising:
a memory that stores a set of instructions; and
at least one processor configured to execute the set of instructions to:
obtain, from one or more databases, construction information about piping, design information about the piping, and material information about the piping;
generate a criteria distribution that is statistical data of criteria data of piping based on the construction information about the piping, the design information about the piping, and the material information about the piping;
generate a changed-state distribution that is statistical data of changed-state data of the piping that have changed due to aging based on at least either vibration or dynamic pressure of the piping that have changed due to aging;
measure, by at least one of an acceleration sensor, a displacement sensor, a velocity sensor, a water pressure meter, a hydrophone, and a dynamic pressure sensor, at least either vibration or dynamic pressure of the piping;
determine deterioration of the piping based on a learned model, and at least either the measured vibration or the measured dynamic pressure of the piping;
update at least either the criteria distribution or the changed-state distribution based on a result of the determination on the deterioration of the piping;
newly measure at least either vibration or dynamic pressure of the piping; and
newly determine deterioration of the piping based on the learned model, and at least either the newly measured vibration or the newly measured dynamic pressure of the piping; wherein,
the learned model is machine-learned based on the criteria distribution or the changed-state distribution.

2. The piping diagnostic device according to claim 1 , wherein
the at least one processor is further configured to:
calculate a discriminant index based on at least either the vibration or the dynamic pressure of the piping;
determine whether or not the discriminant index is abnormal based on the discriminant index and a predetermined threshold value; and
identify, when the discriminant index is determined to be abnormal, a cause thereof, and rewriting at least one of the construction information, the design information, and the material information; and
determine deterioration of the piping, when the discriminant index is determined to be normal.

3. The piping diagnostic device according to claim 2 , wherein
the discriminant index is at least either sound velocity of the vibration or a natural frequency of the piping.

4. The piping diagnostic device according to claim 1 , wherein
the at least one processor is further configured to
use at least one selected from a group including a finite element method and a theoretical expression based on elastic theory, the finite element method using a material constant, a dimensions constant, and a boundary condition as input.

5. The piping diagnostic device according to claim 1 , further comprising
the at least one processor is further configured to:
generate a discriminant criterion as the learned model based on the criteria distribution and the changed-state distribution; and
determine deterioration of the piping based on the discriminant criterion and at least either the vibration or the dynamic pressure of the piping.

6. The piping diagnostic device according to claim 5 , wherein
the at least one processor is further configured to
generate the discriminant criterion by at least one selected from a group including a support vector machine, a linear classifier, a kernel classifier, a k-NN classifier, and a deep neural network.

7. A piping diagnostic method comprising:
obtaining, from one or more databases, construction information about piping, design information about the piping, and material information about the piping;
performing criteria distribution generation including generating a criteria distribution that is statistical data of criteria data of the piping based on the construction information about the piping, the design information about the piping, and the material information about the piping;
generating a changed-state distribution that is statistical data of changed-state data of the piping that have changed due to aging based on at least either vibration or dynamic pressure of the piping that have changed due to aging;
measuring, by at least one of an acceleration sensor, a displacement sensor, a velocity sensor, a water pressure meter, a hydrophone, and a dynamic pressure sensor, at least either vibration or dynamic pressure of the piping;
performing deterioration determination including determining deterioration of the piping based on a learned model, and at least either the measured vibration or the measured dynamic pressure of the piping;
updating at least either the criteria distribution or the changed-state distribution based on a result of the determination on the deterioration of the piping;
newly measuring at least either vibration or dynamic pressure of the piping; and
newly determining deterioration of the piping based on the learned model, and at least either the newly measured vibration or the newly measured dynamic pressure of the piping; wherein,
the learned model is machine-learned based on the criteria distribution or the changed-state distribution.

8. The piping diagnostic method according to claim 7 further comprising:
calculating a discriminant index based on at least either the vibration or the dynamic pressure of the piping;
determining whether or not the discriminant index is abnormal based on the discriminant index and a predetermined threshold value;
identifying, when the discriminant index is determined to be abnormal, a cause thereof and rewriting at least one of the construction information, the design information, and the material information; and
determining deterioration of the piping in the deterioration determination, when the discriminant index is determined to be normal.

9. The piping diagnostic method according to claim 8 , wherein
the discriminant index is at least either sound velocity of the vibration or a natural frequency of the piping.

10. The piping diagnostic method according to claim 7 , comprising
using, for the criteria distribution generation, at least one selected from a group including a finite element method and a theoretical expression based on elastic theory, the finite element method using a material constant, a dimensions constant, and a boundary condition as input.

11. The piping diagnostic method according to claim 7 , further comprising
performing discriminant problem optimization including generating a discriminant criterion as the learned model based on the criteria distribution and the changed-state distribution, wherein
the deterioration determination includes determining deterioration of the piping based on the discriminant criterion and at least either the vibration or the dynamic pressure of the piping.

12. The piping diagnostic method according to claim 11 , wherein
the discriminant problem optimization uses at least one selected from a group including a support vector machine, a linear classifier, a kernel classifier, a k-NN classifier, and a deep neural network.

13. A non-transitory computer-readable storage medium storing a program making a computer execute:
obtainment processing of obtaining, from one or more databases, construction information about piping, design information about the piping, and material information about the piping;
criteria distribution generating processing of generating a criteria distribution that is statistical data of criteria data based on the construction informatio

Assignees

• Nec Corp

Inventors

• Takata Soichiro
• Kumura Takahiro
• Miyake Kanta
• Tominaga Shin
• Inoue Hirofumi
• Shinoda Shigeki

Classifications

• G01M3/2815

  using pressure measurements · CPC title

• G01M3/28 Primary

  for pipes, cables or tubes; for pipe joints or seals; for valves {; for welds} · CPC title

• F17D5/02

  Preventing, monitoring, or locating loss · CPC title

• G01M3/243

  for pipes · CPC title

• F17D5/06 Primary

  using electric or acoustic means · CPC title