The PHM Data Challenge is a competition open to all potential conference attendees. The key dates along with detailed information on this event are now announced. The challenge of this year is focused on tracking the healthy state of components and predicting the date of upcoming gearbox replacement within a pulverizer system in an operating steam based power plant. This is a fully open competition in which collaboration is encouraged. The teams may be composed of any combination of students, researchers, and industry professionals. The results will be evaluated by the PHMAP Local Organizing Committee and all teams will be ranked. The top three scoring teams will be invited to present at a special session of the conference and will be recognized during the conference.
Teams Collaboration is encouraged and teams may be comprised of one of more students and professionals. The team judged to have the first, second, and third best scores will be awarded cash prizes of USD 600, 400, and 200 respectively contingent upon:
Having at least one member of the team attend the PHMAP 2017 Conference.
Presenting the analysis results and technique employed at a special session within the Conference program
Submitting a peer-reviewed conference paper (three to four page-long short paper, Submission of the papers is outside the regular paper submission process and follows its own modified schedule.)
The organizers of the competition reserve the right to both modify these rules and disqualify any team for any practices it deems inconsistent with fair and open practices.
Registration Teams should register by contacting the competition organizers (yhseo@kimm.re.kr, hsoh@gist.ac.kr) with their name(s), a team alias under which the scores would be posted.
* Please note: In the spirit of fair competition, we allow only one account per team. Please do not register multiple times under different user names, under fictitious names, or using anonymous accounts. Competition organizers reserve the right to delete multiple entries from same persons or teams and to disqualify those who are trying to game the system or using fictitious identities.
Key Dates of Data Challenge
Competition Open
Feb. 15, 2017
Competition Closed
Jun. 2, 2017
Preliminary Winner Announcement
Jun. 11, 2017
Winning Paper Submission Due
Jul. 1, 2017
Winner Announcement
Jul. 5, 2017
PHM Conference Dates
July 12, 2017
Submission for Data Challenge Each team is permitted one submission every 2 weeks (detailed dates are listed as below). A correct submission will be composed of a CSV file whose name is the team alias (ex: If the alias is "solution", then the file name will be "solution_team name.csv"). The CSV file should have three columns
- First column - Name of Pulverizer - Second column - Estimated date for gearbox replacement - Third column - Highly risk component of the gearbox
Detailed Dates for Submission
Preliminary (first submission)
Apr. 21, 2017
Mid-term (second submission)
May 5, 2017
Semi-final (third submission)
May 19, 2017
Final (last submission)
Jun. 2, 2017
Scoring
Since the test set contains actual problems of power plant, the deviation (scoring function) from the exact answers will be calculated as below equation.
The error (errorn, where n=1, 2, and 3) of each system is the difference between a predicted replacement date and the actual replacement date for gearboxes (should be suggested as 'day(s)' and the actual date of each system).
If the most severe component (close to failure) of each system is correctly answered, then a weighting factor (wn=0.7) will be multiplied to the error term. Otherwise, another weighting factor (wn=1.0) will be applied.
Note: a, b, and c are predefined constants.
Data A pulverizeris a grinding system of coal for boiler in steam generating power cycle. More than 20 years ago, six pulverizers (A-F) were installed and have been operated. To operate the pulverizer, a motor-driven gearbox is connected to a large size grinding system.
For training, sensor data from the pulverizers C, D, and F are provided. They were collected for a given period of time. The gearboxes in the pulverizers performed as intended without any failure during that period of operation, although there may be normal degradation.
For the pulverizers, A, B and E, predict the replacement date (day/month/year) of the gearboxes and find the most severely-degraded component (shown in Fig. 3) of the gearboxes close to failure of each pulverizer. Detailed operating information of the pulverizers and gearboxes are given here.
The description of data file is explained below:
In the given data set, file name is recorded in the form of [Pulv_X_ Y- Measuring date (Year-Month-Day) - Measuring time (Hour-Min-Sec), where X: Pulverizer A-E and Y: Measuring point-Body Case-1, Body Case-2, MTR-I_B and MTR-O_B]
< Training data set >
- Pulverizer C : data recorded from 2011/06/14 (year/month/day) to 2013/06/12 - Pulverizer D : data recorded from 2011/06/14 to 2013/06/12 - Pulverizer F : data recorded from 2011/06/14 to 2013/06/12
< Test data set >
- Pulverizer A : data recorded from 2011/06/14 (year/month/day) to 2012/07/31 - Pulverizer B : data recorded from 2011/09/08 to 2013/06/12 - Pulverizer E : data recorded from 2011/06/14 to 2013/07/31
(* Note for the pulverizer E: the data recorded on 2012/07/12 includes the increase of motor current due to the stuck bowl roll during operation. Therefore, roll assembly part was only replaced with new one. After replacement, all data shows a stable condition. For operation, the data recorded at 2012/09/24 and 2012/09/26 is for the idle and load conditions, respectively.)
< Failure set >
- Failure 1 : data recorded 81 days earlier than the gearbox failure - Failure 2 : data recorded 1 day earlier than the gearbox failure
* To download all the data above, please click here.