Skip to Main Content
 

Global Search Box

 
 
 
 

ETD Abstract Container

Abstract Header

KNOWLEDGE DISCOVERY USING DATA ANALYSIS TECHNIQUES AND INVERSE EXTREME VALUE STATISTICS TO BETTER PREDICT LIFE OF A BEARING

Abstract Details

2013, Master of Science, Ohio State University, Industrial and Systems Engineering.
Predicting bearing life is an important task for bearing companies because it is vital in the eyes of the customer. Testing methods have improved a lot with the advancements in technology. Predicting the bearing life accurately and creating models which provide results very close to the actual measured values provided by these advanced methods thus becomes a difficult and important challenge for many manufacturing companies. Here, we focus on cases in which there is an ability to collect special types of data. In particular, data can be collected about the inclusion size and tested life from a large number of samples having the same material, from the same plant, of the same design, and from the same process. Further, with respect to predicting the future life of a given bearing, there is an ability to measure the maximum inclusion size from a sample of material from the same batch of material. Under all these conditions, the thesis proposes a highly accurate estimation procedure. A model is created using concepts from literature, which discuss about the inclusions and their effect on bearing life, Rolling Contact Fatigue (RCF) mechanisms and Extreme value statistics. The proposed procedure involves the following steps - 1. Filter the data based on the parameters i.e. same plant, process, design and material quality 2. Fit an extreme value distribution to the set of data collected on life and compute of the distribution parameters. 3. Using these parameters and performing some transformations a plot is constructed. Now this model represents bearings population which have the same set of parameters as mentioned in step 1 4. Data on inclusions is collected from the materials that are used in the manufacturing of the remainder of the bearings which have the same set of parameters. 5. Using the plot we can compute the life of the bearing. By using the method described above a model was created which accurately predicts the life. Upon carrying out a verification process with new data sets showed that the new model is able to predict life values which are much closer to the measured values, when compared to the current model being used.
Theodore Allen, PhD (Advisor)
Rajiv Shivpuri, PhD (Other)

Recommended Citations

Citations

  • Hari, R. (2013). KNOWLEDGE DISCOVERY USING DATA ANALYSIS TECHNIQUES AND INVERSE EXTREME VALUE STATISTICS TO BETTER PREDICT LIFE OF A BEARING [Master's thesis, Ohio State University]. OhioLINK Electronic Theses and Dissertations Center. http://rave.ohiolink.edu/etdc/view?acc_num=osu1366304299

    APA Style (7th edition)

  • Hari, Rohit. KNOWLEDGE DISCOVERY USING DATA ANALYSIS TECHNIQUES AND INVERSE EXTREME VALUE STATISTICS TO BETTER PREDICT LIFE OF A BEARING. 2013. Ohio State University, Master's thesis. OhioLINK Electronic Theses and Dissertations Center, http://rave.ohiolink.edu/etdc/view?acc_num=osu1366304299.

    MLA Style (8th edition)

  • Hari, Rohit. "KNOWLEDGE DISCOVERY USING DATA ANALYSIS TECHNIQUES AND INVERSE EXTREME VALUE STATISTICS TO BETTER PREDICT LIFE OF A BEARING." Master's thesis, Ohio State University, 2013. http://rave.ohiolink.edu/etdc/view?acc_num=osu1366304299

    Chicago Manual of Style (17th edition)