Select a calendar:
Filter July Events by Event Type:
SUNMONTUEWEDTHUFRISAT
Events for July 22, 2013
-
Meet USC: Admission Presentation, Campus Tour, & Engineering TalkMon, Jul 22, 2013
Viterbi School of Engineering Undergraduate Admission
Receptions & Special Events
This half day program is designed for prospective freshmen and family members. Meet USC includes an information session on the University and the Admission process; a student led walking tour of campus and a meeting with us in the Viterbi School. Meet USC is designed to answer all of your questions about USC, the application process and financial aid. Reservations are required for Meet USC. This program occurs twice, once at 8:30 a.m. and again at 12:30 p.m. Please visit https://esdweb.esd.usc.edu/unresrsvp/MeetUSC.aspx to check availability and make an appointment. Be sure to list an Engineering major as your "intended major" on the webform!
Location: Ronald Tutor Campus Center (TCC) - USC Admission Office
Audiences: Everyone Is Invited
Contact: Viterbi Admission
-
PhD Defense - Yintao Liu
Mon, Jul 22, 2013 @ 01:00 PM - 03:00 PM
Thomas Lord Department of Computer Science
University Calendar
PhD Candidate: Yintao Liu
Committee members:
Cauligi S. Raghavendra (chair)
Ke-Thia Yao (co-chair)
Aiichiro Nakano
Iraj Ershaghi (outside member)
Time: Jul 22 1pm-3pm
Location: RTH 324
Title: Failure Prediction for Rod Pump Artificial Lift Systems
Abstract:
Failure predictions, a subset of anomaly detection which aims at the precursory events that potentially triggers failures, is of greater value in real life. Given massive amount of historical data in multivariate time series, data mining techniques can play an important role that learns from historical failures which can then be integrated in real world applications. However because of the rarity of such events with regards to thousands of assets and diversity of failure patterns, it is unrealistic to build individual predictive model for each asset. Moreover, uncertainties such as missing data, noise, data inconsistency and approach to combine domain knowledge makes it an even more challenging problem. This thesis addresses the problems of failure prediction on multiple multivariate time series: 1) how to systematically learn from historical failures and train an effective model that is applicable in failure prediction application; 2) how to train a generalized model from the labeled dataset that is efficient in predicting failures out of thousands of multivariate time series that exhibit clustered heterogeneity. This thesis emphasizes, but not limited to, down-hole mechanical failures for rod pump artificial lift systems, which is the most common type of oil producer systems.
The first part of the thesis presents Smart Engineering Apprentice system (SEA) that involves data extraction, data preparation, feature extraction, data mining, alert generation and knowledge management. The data extraction stage extracts data needed including the time series data and event logs from the enterprise database. The noise and missing values are partly handled in the data preparation stage. The denoised data is then fed into feature extraction stage for obtaining features. Given a labeled dataset, general supervised learning algorithms can be applied to train, test and evaluate the results in the Data Mining stage. In the Alerting stage the system visually depicts alerts to provide warning of impending failures. Finally, in knowledge management stage, a wide range of factors are combined to train a confidence level model for ranking across multiple assets from multiple categories. SEA has been successfully applied to real-world failure prediction for rod pump artificial lift systems.
In the second part, this thesis presents an in-depth model for generalizing the learning algorithm so that a unified model can be applied to multiple heterogeneous fields yet maintaining comparable precision and recall. Our objective was to build such a generalized model that: 1) automatically recognizes the easy examples based on the limited knowledge from the subject matter experts (SME); 2) takes advantage of larger amount of recognizable examples from all historical data so that the learned model is more statistically robust; 3) better customizes the model such that different oil fields are capable of exhibiting variations that arises from other important uncertainties that were difficult to be considered during previous algorithms. We proposed an unsupervised rule-enhanced labeling with support vector machines that enables the SEA system to learn from much larger historical data from multiple fields. Then we further improved this algorithm by proposing a multitask learning algorithm that combines multiple decision relevant factors to yield a better generalized global model.
SEA system is evaluated using real-world data from rod pump artificial lift systems with good results and significant economic value for use in oil fields.
Location: Ronald Tutor Hall of Engineering (RTH) - 324
Audiences: Everyone Is Invited
Contact: Lizsl De Leon
