Cyber-Physical System Augmented Prognostics and Health Management for Fleet-Based Systems

Cyber-Physical System Augmented Prognostics and Health Management for Fleet-Based Systems

Bibliographic Details
Main Author: Liu, Zongchang
Language:English
Published: University of Cincinnati / OhioLINK 2018
Subjects:
Mechanical Engineering
Prognostics and health management
Cyber-physical systems
networked and fleet-sourced systems
data-driven PHM
Online Access:http://rave.ohiolink.edu/etdc/view?acc_num=ucin1522321192371536
id ndltd-OhioLink-oai-etd.ohiolink.edu-ucin1522321192371536
record_format oai_dc
collection NDLTD
language English
sources NDLTD
topic Mechanical Engineering
Prognostics and health management
Cyber-physical systems
networked and fleet-sourced systems
data-driven PHM
spellingShingle Mechanical Engineering
Prognostics and health management
Cyber-physical systems
networked and fleet-sourced systems
data-driven PHM
Liu, Zongchang
Cyber-Physical System Augmented Prognostics and Health Management for Fleet-Based Systems
author Liu, Zongchang
author_facet Liu, Zongchang
author_sort Liu, Zongchang
title Cyber-Physical System Augmented Prognostics and Health Management for Fleet-Based Systems
title_short Cyber-Physical System Augmented Prognostics and Health Management for Fleet-Based Systems
title_full Cyber-Physical System Augmented Prognostics and Health Management for Fleet-Based Systems
title_fullStr Cyber-Physical System Augmented Prognostics and Health Management for Fleet-Based Systems
title_full_unstemmed Cyber-Physical System Augmented Prognostics and Health Management for Fleet-Based Systems
title_sort cyber-physical system augmented prognostics and health management for fleet-based systems
publisher University of Cincinnati / OhioLINK
publishDate 2018
url http://rave.ohiolink.edu/etdc/view?acc_num=ucin1522321192371536
work_keys_str_mv AT liuzongchang cyberphysicalsystemaugmentedprognosticsandhealthmanagementforfleetbasedsystems
_version_ 1719453494009659392
spelling ndltd-OhioLink-oai-etd.ohiolink.edu-ucin15223211923715362021-08-03T07:05:42Z Cyber-Physical System Augmented Prognostics and Health Management for Fleet-Based Systems Liu, Zongchang Mechanical Engineering Prognostics and health management Cyber-physical systems networked and fleet-sourced systems data-driven PHM Owning to the increasing demand of asset reliability and operation efficiency, the research field of prognostics and health management (PHM) is gaining more motivation and efforts from both academy and industry. The booming development of industrial Internet of Things (IIoT) has made data collection seamless, and is creating the unprecedented big data environment that brings both opportunity and challenges to operation optimization of industrial assets. Despites fast advancements of data-driven PHM techniques, the application of conventional data abstraction and pattern recognition methods face many challenges when dealing with big volume, variation, and veracity data environment. The challenges of data-driven PHM come from the following aspects. To begin with, the conventional `Train-Test-Implement’ process for model development relies on the comprehensiveness and supervised labeling of historical data of each individual asset. This process is difficult to be scalable when the data analyst is dealing with large volume of assets and data. Secondly, the accuracy of data-driven fault diagnosis and prognosis model relies on the quality of training data, and won’t adapt itself when encounter data from an unknown status. When coming up with such situation, the model has to be retrained with data containing the new status. The third challenge is to convert the prognosis results into actionable insights for optimized operation plans. The most popular output metrics of data-driven PHM models are virtual indexes, such as distance, likelihood, residual, etc., which can hardly be inferred to operation decisions. The referential information from prognosis model is desired to contain existing failure mode, root cause, deterioration process, assessment of impact, and future development trend, which can be used as input to decision-making models to generate the optimal operation plan. The concept of Cyber-Physical Systems (CPS) has been defined as an engineered system in which the nature and human made systems are integrated with computation, communication, and control systems in all scales. A `5C’ architecture for CPS-enabled smart maintenance, which consists of `Connection’, `data-to-information Conversion’, `Cyber’, `Cognition’, and `Configuration’, has been proposed by Lee, et al [1], [2]. This dissertation work is aimed to incorporate the `5C’ architecture of Cyber-physical system in the development and implementation of data-driven PHM. It highlights the technical framework in the `Cyber’ and `Cognition’ level, and provides guidance on the interaction mechanism between `Conversation’, `Cyber’, and `Cognation’ level to enable self-adaptive and self-resilience of PHM system. The dissertation further implements the proposed framework on two case studies. The first case study is on remaining useful life (RUL) prediction of aircraft engines, with demonstration of using peer-to-peer prognosis techniques to build library of degradation trajectories with historical data from a fleet of engines, and predict the RUL of testing data set engines with real-time data stream. The second is on predictive maintenance of wind turbines. Demonstration of utilizing turbine fleet data for health assessment, fault diagnosis, root cause identification, impact assessment, and maintenance plan optimization will be presented and benchmarked with existing methods. The figure of merit of the proposed framework will be measured on the maintenance cost reduction from historical data of real-world wind farms. 2018-05-15 English text University of Cincinnati / OhioLINK http://rave.ohiolink.edu/etdc/view?acc_num=ucin1522321192371536 http://rave.ohiolink.edu/etdc/view?acc_num=ucin1522321192371536 unrestricted This thesis or dissertation is protected by copyright: some rights reserved. It is licensed for use under a Creative Commons license. Specific terms and permissions are available from this document's record in the OhioLINK ETD Center.

Similar Items