1. Introduction

Japan Aerospace Exploration Agency (JAXA) aims to improve PHM technology for spacecraft propulsion systems. Telemetry data that can be acquired in orbit is limited due to the limitation of sensor installations and downlink capacity. A numerical simulator to predict the dynamic response of a spacecraft propulsion system with high accuracy has been studied to generate a dataset covering normal conditions and all anticipated fault scenarios in actual equipment.

PHMAP 2023 Data Challenge Leaderboard

2. Competition Objectives

Improve PHM technology for next-generation spacecraft by diagnosing normal, bubble anomalies, solenoid valve faults, and unknown abnormal cases using data generated by a simulator for a simplified propulsion system developed with the cooperation of JAXA.

3. Experiment Scenarios

Figure 1 shows an experimental propulsion system. The working fluid is water pressurized to 2 MPa and discharged through four solenoid valves (SV1 – SV4) simulating thrusters. P1 – P8 shows the pressure sensor and time series data is obtained at a sampling rate of 1 kHz from 0-1200 ms. By opening and closing the solenoid valves, pressure fluctuation is observed due to the water hammer followed by acoustic modes inside the propulsion system.
Typical time series data is shown in Fig. 2. The solenoid valve opens at 100 ms and closes at 300 ms. To account for individual differences of solenoid valve appearing in actual equipment, the valve movement has an uncertainty of 1 ms. The opening and closing times remain at 400 ms, even with the uncertainty (e.g., open for 99.7 ms and close for 300.3 ms). This sequence is conducted three times successively, resulting in a total measurement of 1200 ms.

Fig. 1 Schematic of experimental propulsion system.

Fig. 2 Typical pressure profile.

In this competition, anomalies due to bubble contamination and fault due to the abnormal opening of the solenoid valve are considered. Moreover, an unknown anomaly is included in the test data.

Bubble anomaly:

Air bubbles occasionally appear in the pipes during the actual operation of a spacecraft. The existence of bubbles changes the speed of sound, causing slight changes in the pressure fluctuations. It is desirable to detect the appearance of bubble and their location.

• There are eight possible locations, BV1 and BP1 to BP7 as shown in Fig.1.
• The amount of bubbles contaminated in the propulsion system is constant in all cases for simplicity.

Solenoid valve faults:

This is one of the major failure modes in spacecraft propulsion systems. It is required to find which solenoid valves failed and their opening ratio.

• Solenoid valves open and close at the opening ratio of 100% and 0%, respectively.
• In case of faults, solenoid valves open at a degree between 0% and 100%, which results in a reduced volume of fluid through the solenoid valve.

Unknown anomaly:

In the practical operation, completely unforeseen and unknown anomalies or faults may occur. It is also required to distinguish unknown anomalies without confusing them with known anomalies and faults. Some unknown anomalies or faults are mixed in the test data. Identifying them is also part of the task in this competition.

Individual differences in spacecraft:

Since solenoid valves have individual differences, such as the timing of opening and closing, time series data acquired from the spacecraft propulsion system show differences, which in turn lead to individual differences in the spacecraft. In this competition, four spacecraft (No.1 to 4) are considered. Results for the three of them, No.1 to No.3, are included in the training data, but test data is composed of the results for No.1 and No.4.

4. Prediction Goals

1. Determine normal or abnormal for all test data.
   2. For the data detected as abnormal, determine if it is an anomaly due to bubble contamination, solenoid valve fault, or unknown fault.
   3. For the data identified as bubble contamination, determine the location of the bubble from eight locations, BV1, and BP1 to BP7.
   4. For the data identified as solenoid valve fault, determine which of the four solenoid valves (SV1 to SV4) failed.
   5. For the solenoid valve identified as a fault, predict the opening ratio. (0% <= Opening ration < 100%)

5. Dataset

You can download the dataset here.

Teams

Collaboration is encouraged and teams may be comprised of one or more students and professionals. The teams judged to have the first, second, and third best scores will be awarded prizes, contingent upon:

• Having at least one member of the team register and attend the PHM 2022 Conference.
• Submitting a journal paper for the IJPHM journal.
• Presenting the analysis results and technique employed at a special session within the conference program.

The organizers of the competition reserve the right to both modify these rules and disqualify any team for any efforts it deems inconsistent with fair and open practices.

Data Challenge Registration

Teams may register here. Please note that the registration is a 2 steps process:
1) You must first log in with a PHM user account to complete the form. If you need to create one here –> Sign Up
2) Once logged in, you need to complete the Data Challenge Application form. Then, we will grant your user access to use the Data Challenge submission area.

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 the same person (or team) and/or to disqualify those who are trying to “game” the system or using fictitious identities.

Relevant Dates

Data Challenge Details

The details about this year’s data challenge can be found in this document. For each CSV file that is provided, the first column is the label for that cycle, in which a label of 1 represents the no-fault condition, a label of 2-11 correspond to different fault types.  A label of 0, which is present for the testing data for the online leaderboard and the final validation data set, indicate an unknown/blind label, in which the objective is to predict the label for each cycle. The columns of data after the label represent the pressure signal data for that cycle.

Dataset

Training and Testing Dataset

The training and test datasets are available here: Training Data Set,  Testing Data Set

Validation Dataset

The one-time validation data set is available here: Validation Data Set
NOTE:  The official scoring for the competition is based solely on the validation data set.

Data Challenge Submission

For Testing Dataset:

Please upload your submission in the Data Challenge submission area. Note that you will only see the Submission area if you have previously applied using the application form and we have granted you access

Please ensure the filename is yourusername.txt or else the automated scorer will not be able to read your file and score your algorithm and the conference organizers will not be able to consider your submission as a valid submission. Please ensure that your results are in the format shown in this Data_Challenge_PHM2022_Example_Submission.txt file which contains only 3184 values in a column vector format.

Also, you only can submit (or overwrite) the yourusername.txt file for your submission as often as you’d like.  The system will automatically score the most current submission twice daily, at noon and midnight US Pacific Daylight Time (0700 and 2100 UTC).  You may continue to upload submission of testing data to the submission area after the validation data is available, and they will continue to be scored and displayed.

For Validation Dataset:

Please upload your submission in the Data Challenge Validation submission area. Note that you will only see the Submission area if you have previously applied using the application form and we have granted you access

Please ensure the filename is yourusername.txt or else the automated scorer will not be able to read your file and score your algorithm and the conference organizers will not be able to consider your submission as a valid submission.  Since the validation data consists of two individual configurations (#7 and #8), please combine the results into a single file.  The first 7935 rows should correspond to #7 and the last 8461 rows corresponding to #8, for a total of 16396 rows.  A format checker has been implemented that displays (in no particular order, other than alphabetically) a table of all current entries and whether or not the system recognizes it as correctly formatted.  This checker runs at 30 minutes after each hour.

Also, you only can submit (or overwrite) the yourusername.txt file for your validation submission as often as you’d like.  However, only the most current submission at 23:59:59 US PDT on August 12, 2022 (August 13, 2022, 06:59:59 UTC), will be scored for the competition.