Automated Fault Diagnosis of a Micro Turbine With Comparison to a Neural Network Technique
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
In the predicted future of distributed power generation, a large number of users will operate gas turbine powered cogeneration systems. These systems will be small, relatively inexpensive, and installed in locations without ready access to experts in gas turbine maintenance. Consequently, an automated system to monitor the engine and diagnose the health of the system is required. To remain compatible with the low cost of the overall system, the diagnostic system must also be relatively inexpensive to install and operate. Therefore, a minimum number of extra sensors and computing power should be used. A statistical technique is presented that compares the engine operation over time to the expected trends for particular faults. The technique ranks the probability that each fault is occurring on the engine. The technique can be used online, with daily data from the engine forming a trend for comparison, or, with less accuracy, based on a single operating point. The use of transient operating data with this technique is also examined. This technique has the advantage of providing an automated numerical result of the probability of a particular mode of degradation occurring, but can also produce visual plots of the engine operation. This allows maintenance staff to remain involved in the process, if they wish, rather than the system operating purely as a black box, and provides an easy to understand aid for discussions with operators. The technique is compared to an off the shelf neural network to determine its usefulness in comparison to other diagnostic methods. The test bed was a micro turbojet engine. The data to test the system was obtained from both experiment and computer modeling of the test engine.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.001 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it