Pump assembly health assessment

Patent metadata

Field	Value
Publication number	US-10584698-B2
Application number	US-201615092694-A
Country	US
Kind code	B2
Filing date	Apr 7, 2016
Priority date	Apr 7, 2016
Publication date	Mar 10, 2020
Grant date	Mar 10, 2020

Abstract

Method and apparatus for assessing health of fracturing fluid pump assemblies and other wellsite equipment. For example, predicted data indicative of a first operational parameter of a pump assembly is generated utilizing: a model relating the first operational parameter to each of a plurality of second operational parameters of the pump assembly; and real-time data indicative of each of the second operational parameters. Health of the pump assembly is then assessed based on: the predicted data indicative of the first operational parameter; and real-time data indicative of the first operational parameter.

First claim

What is claimed is:

1. A method comprising:
operating a processing system comprising a processor and a memory including computer program code, wherein operating the processing system comprises:
operating a wellsite surface equipment pump assembly, wherein the pump assembly comprises a plurality of components and is disposed at a wellsite surface from which a wellbore originates and extends below to a subterranean formation, and wherein the pump assembly components comprise a prime mover, a transmission, a fluid end, and a power end;
generating predicted data indicative of a first operational parameter of at least one component of the pump assembly, wherein generating the predicted data utilizes:
a model relating the first operational parameter to each of a plurality of second operational parameters of the components of the pump assembly; and
real-time operational data indicative of each of the second operational parameters; and

assessing health of at least one of the components of the pump assembly based on:
the predicted data indicative of the first operational parameter; and
real-time operational data indicative of the first operational parameter; and

generating a health index of at least one of the components of the pump assembly based on a relationship between the predicted and real-time operational data indicative of the first operational parameter.

2. The method of claim 1 wherein the health assessment comprises:
generating residual data comprising quantitative differences between temporally-related ones of the predicted and real-time data indicative of the first operational parameter; and
comparing the residual data to a predetermined threshold.

3. The method of claim 2 wherein the pump assembly is a first one of a plurality of pump assemblies each having similar pump assembly components and disposed at the wellsite surface, and wherein the predetermined threshold corresponds to a known level of degradation of one or more second ones of the pump assemblies.

4. The method of claim 3 wherein:
comparing the residual data to a predetermined threshold comprises comparing the residual data to a first predetermined threshold and a second predetermined threshold;
the first predetermined threshold corresponds to a first predetermined value related to operational failure of one or more of the second ones of the pump assemblies; and
the second predetermined threshold is less than the first predetermined threshold and corresponds to reduced operational effectiveness of one or more of the second ones of the pump assemblies.

5. The method of claim 1 wherein the processing system, the processor, and the memory are a first processing system, a first processor, and a first memory, respectively, wherein the first processing system is separate and discrete from a second processing system comprising a second processor and a second memory, and wherein the method further comprises operating one of the first and second processing systems to generate the model.

6. The method of claim 5 wherein:
the pump assembly is a first pump assembly;
a plurality of pump assemblies, each having similar pump assembly components, comprises the first pump assembly and a plurality of second pump assemblies;
generating the model utilizes preexisting training data indicative of each of the first and second operational parameters of at least each of the second pump assemblies; and
generating the model comprises iteratively adjusting a plurality of fitting parameters of the model to exploit a correlation between each of the first and second operational parameters in the preexisting training data to optimize prediction performance of the model.

7. The method of claim 6 wherein the preexisting training data utilized to generate the model is indicative of each of the first and second operational parameters of each of the first and second pump assemblies.

8. The method of claim 5 wherein:
the pump assembly is a first pump assembly;
a plurality of pump assemblies, each having similar structure and function, comprises the first pump assembly and a plurality of second pump assemblies;
generating the model comprises:
accessing preexisting training data indicative of each of a plurality of available operational parameters of at least each of the second pump assemblies; and
selecting the first and second operational parameters from the plurality of available operational parameters based on correlation between operational degradation of ones of at least the second pump assemblies and ones of the plurality of available operational parameters.

9. The method of claim 8 wherein selecting the first and second operational parameters comprises:
generating a plurality of feature-selection models each predicting a corresponding one of the plurality of available operational parameters utilizing other ones of the plurality of available operational parameters; and
selecting the first and second operational parameters based on correlation of each of the plurality of feature-selection models.

10. An apparatus comprising:
a processing system comprising a processor and a memory including computer program code, wherein the processing system:
gathers data from a plurality of operating components of at least one wellsite surface equipment pump assembly, wherein the wellsite surface equipment pump assembly components comprise at least a prime mover, a transmission, a fluid end, and a power end, and wherein the wellsite surface equipment pump assembly and each of its operating components are disposed at a wellsite surface from which a wellbore originates and extends below to a subterranean formation;
generates predicted data indicative of a first operational parameter of the plurality of operating components of the at least one wellsite surface equipment pump assembly, wherein the processing system generates the predicted data utilizing:
a model implemented in the computer program code and relating the first operational parameter to each of a plurality of second operational parameters of the wellsite surface equipment pump assembly; and
real-time data indicative of each of the second operational parameters; and

assesses, without human interaction, health of at least one of the operating components of the wellsite surface equipment pump assembly based on:
the predicted data indicative of the first operational parameter; and
real-time data indicative of the first operational parameter.

11. The apparatus of claim 10 wherein the processing system assesses the health by:
generating residual data comprising quantitative differences between temporally-related ones of the predicted and real-time data indicative of the first operational parameter; and
comparing the residual data to a predetermined threshold.

12. The apparatus of claim 11 wherein the pump assembly is a first one of a plurality of pump assemblies each having similar pump assembly components, and wherein the predetermined threshold corresponds to a known level of degradation of one or more second ones of the pump assemblies.

13. The apparatus of claim 12 wherein:
comparing the residual data to a predetermined threshold comprises comparing the residual data to a first predetermined threshold and a second predetermined threshold;
the first predetermined threshold corresponds to a first predetermined value related to operational failure of one or more of the second ones of the pump assemblies; and
the second predetermined threshold is less than the first predetermined threshold and corresponds to reduced operational effectiveness of one or more of the second ones of the pump assemblies.

Assignees

• Schlumberger Technology Corp

Inventors

• Haddad Gilbert
• Williard Nicholas Dane
• Eklund Neil Holger White
• Follini Jean-Marc
• Jaggi Rakesh
• Wolski Shaun Alan
• Chavero Perez Christian Abel
• Zhao Wenyu

Classifications

• F04B1/04

  having cylinders in star- or fan-arrangement · CPC title

• F04B49/065

  and making use of computers · CPC title

• F04B51/00 Primary

  Testing machines, pumps, or pumping installations · CPC title

• F04B53/14

  Pistons, piston-rods or piston-rod connections · CPC title

• F04B43/026

  each plate-like pumping flexible member working in its own pumping chamber · CPC title