From Prognostics and Health Systems Management to Predictive Maintenance 1 – Monitoring and Prognostics

<p>Introduction&nbsp; ix</p>
<p>Chapter 1. PHM and Predictive Maintenance&nbsp; 1</p>
<p>1.1. Anticipative maintenance and prognostics 1</p>
<p>1.1.1. New challenges and evolution of the maintenance function 1</p>
<p>1.1.2. Towards an anticipation of failure mechanisms 3</p>
<p>1.2. Prognostics and estimation of the remaining useful life (RUL)&nbsp; 5</p>
<p>1.2.1. What is it? Definition and measures of prognostics&nbsp; 5</p>
<p>1.2.2. How? Prognostic approaches 6</p>
<p>1.3. From data to decisions: the PHM process&nbsp; 9</p>
<p>1.3.1. Detection, diagnostics and prognostics 9</p>
<p>1.3.2. CBM Architecture and PHM process&nbsp; 10</p>
<p>1.4. Scope of the book 12</p>
<p>Chapter 2. Acquisition: From System to Data&nbsp; 15</p>
<p>2.1. Motivation and content&nbsp; 15</p>
<p>2.2. Critical components and physical parameters&nbsp; 16</p>
<p>2.2.1. Choice of critical components general approach 16</p>
<p>2.2.2. Dependability analysis of the system and related tools&nbsp; 17</p>
<p>2.2.3. Physical parameters to be observed 19</p>
<p>2.3. Data acquisition and storage 20</p>
<p>2.3.1. Choice of sensors 22</p>
<p>2.3.2. Data acquisition&nbsp; 23</p>
<p>2.3.3. Preprocessing and data storage&nbsp; 24</p>
<p>2.4. Case study: toward the PHM of bearings&nbsp; 25</p>
<p>2.4.1. From the train system to the critical component bearing 25</p>
<p>2.4.2. Experimental platform Pronostia 26</p>
<p>2.4.3. Examples of obtained signals 30</p>
<p>2.5. Partial synthesis&nbsp; 30</p>
<p>Chapter 3. Processing: From Data to Health Indicators 33</p>
<p>3.1. Motivation and content&nbsp; 33</p>
<p>3.2. Feature extraction 35</p>
<p>3.2.1. Mapping approaches 35</p>
<p>3.2.2. Temporal and frequency features&nbsp; 36</p>
<p>3.2.3. Time frequency features 38</p>
<p>3.3. Feature reduction/selection&nbsp; 48</p>
<p>3.3.1. Reduction of the feature space&nbsp; 48</p>
<p>3.3.2. Feature selection . 54</p>
<p>3.4. Construction of health indicators 62</p>
<p>3.4.1. An approach based on the Hilbert–Huang transform 62</p>
<p>3.4.2. Approach description and illustrative elements&nbsp; 62</p>
<p>3.5. Partial synthesis&nbsp; 63</p>
<p>Chapter 4. Health Assessment, Prognostics and Remaining Useful Life Part A 67</p>
<p>4.1. Motivation and content&nbsp; 67</p>
<p>4.2. Features prediction by means of connectionist networks 69</p>
<p>4.2.1. Long–term connectionist predictive systems&nbsp; 69</p>
<p>4.2.2. Prediction by means of fast neural networks 77</p>
<p>4.2.3. Applications in PHM problems and discussion 84</p>
<p>4.3. Classification of states and RUL estimation 88</p>
<p>4.3.1. Health state assessment without a priori information about the data 88</p>
<p>4.3.2. Toward increased performances: S–MEFC algorithm 93</p>
<p>4.3.3. Dynamic thresholding procedure&nbsp; 95</p>
<p>4.4. Application and discussion&nbsp; 97</p>
<p>4.4.1. Tests data and protocol&nbsp; 97</p>
<p>4.4.2. Illustration of the dynamic thresholding procedure&nbsp; 101</p>
<p>4.4.3. Performances of the approach&nbsp; 104</p>
<p>4.5. Partial synthesis&nbsp; 105</p>
<p>Chapter 5. Health Assessment, Prognostics, and Remaining Useful Life Part B 109</p>
<p>5.1. Motivation and object 109</p>
<p>5.2. Modeling and estimation of the health state 111</p>
<p>5.2.1. Fundamentals: the Hidden Markov Models (HMM)&nbsp; 111</p>
<p>5.2.2. Extension: mixture of Gaussians HMMs&nbsp; 117</p>
<p>5.2.3. State estimation by means of Dynamic Bayesian Networks 118</p>
<p>5.3. Behavior prediction and RUL estimation&nbsp; 124</p>
<p>5.3.1. Approach: Prognostics by means of DBNs 124</p>
<p>5.3.2. Learning of state sequences 124</p>
<p>5.3.3. Health state detection and RUL estimation 126</p>
<p>5.4. Application and discussion&nbsp; 129</p>
<p>5.4.1. Data and protocol of the tests 129</p>
<p>5.4.2. Health state identification 131</p>
<p>5.4.3. RUL estimation&nbsp; 133</p>
<p>5.5. Partial synthesis&nbsp; 135</p>
<p>Conclusion and Open Issues&nbsp; 137</p>
<p>Bibliography 143</p>
<p>Index&nbsp; 163</p>