• Prognostics & Health Management (PHM)>
• Statistical Model Verification & Validation (V&V)>
• Energy Harvesting (EH)>
• Reliability-based Design Optimization (RBDO)>

Introduction

Prognostics and Health Management (PHM) is a key technology to evaluate the current health condition (health monitoring) and predict the future degradation behavior (health prognostics) of an engineered system throughout its lifecycle.

Sensing FunctionTo ensure high damage detectability by designing an optimal Sensor Network

Reasoning Function

To extract system health relevant information in real-time and to classify system health condition

Prognostics Function

To predict remaining useful lives (RULs) of engineered systems in real-time

Health Management

To enable optimal decision making on maintenance of engineered systems

Health Sensing Function

Objective

To Sense the signal of various system in a cost-effective and high detectable way

Sensor Selection

SensorType

Thermal Sensor

AcousticEmissionSensor

Acc.Sensor

Strain Sensor

Current VoltageSensor O

ilSensor

LoadSensor

Flow Sensor

Appli-cations

Electronics,

Bearings Bearings,

Gearboxes,Engines Bearings,

Gearboxes,

Engines Blade Bearings Engine,

Switches,

Cables

Bearings,

Gearboxes,

Engines

Rotor,

Dams,

Bridges

Hydraulic Component

Sensor Network (SN) Design

To optimize sensor position

To optimize sensor direction

To minimize No. of sensors

While maximizing detectability1

1Ability to detect the health state

Ex) Power transformer

Before SN design

After SN design

Detectability

No. of sensors

Wang, Pingfeng, et al. "A probabilistic detectability-based sensor network design method for system health monitoring and prognostics." Journal of Intelligent Material Systems and Structures (2014): 1045389X14541496.

Health Reasoning Function

Objective

To extract system health relevant information in real-time and to classify system health state

Signal Processing

To enhance detectability of health relevant signals by minimizing obstacles (e.g. noise)

Noisy signal Health relevant signal

Statistical data modeling

To define quantified values representing health state of the system

Health data (HD)

(Multi-variables)

Statistical moments(e.g. RMS)

Model parameters

Health Index (HI)

(Unified variable)

CAE for Health reasoning

To simulate system failure condition by CAE

Hu, Chao, et al. "Copula-based statistical health grade system against mechanical faults of power transformers." Power Delivery, IEEE Transactions on 27.4 (2012): 1809-1819.

Health Prognostics Function

Objective

To define Health Index (HI) and to predict remaining useful lives (RULs) of engineered systems in real-time

Model-based Prognostics

Assumption: Known PoF; few run-to-failure data

Key element: Online identification of PoF-based degradation model

Data-driven Prognostics

Assumption: Massive run-to-failure data

Key element: Offline training and online prediction processes

Loading Signals

Simulation

Identify Model

Simulate with Loading

Response Signals

Estimation

Update Parameters

Update & Project HI

Predicted RUL

TestingSignals

Online Process

Extract Online HI

Project or Interpolate

Training Signals

Offline Process

Extract Offline HI

Build Health Knowledge

Hu, Chao, et al. "Ensemble of data-driven prognostic algorithms for robust prediction of remaining useful life." Reliability Engineering & System Safety103 (2012): 120-135.

Application: Smart Factory

SMART FACTORY

Equipment

Health

Product Health

Equipment Health Management

Health Prognosis of Product & Equipment

Integrated Health Management System for Product & Equipment

Big Data

Deep Learning

Domain Knowledge

Integrated Health Prediction/Management System for Product & Manufacturing Equipment