We Released a Motor Current Signature Analysis (MCSA) Model: MantisV2 at F1 0.943, and What That Number Does Not Prove
We released a research model on Hugging Face that classifies a motor’s fault state from its three-phase current alone. It is called Thaki-MCSA-MantisV2-MCC5. We fine-tuned the pretrained time-series model MantisV2 on the MCC5 motor current data, and it distinguishes 24 states.
We evaluated it once, on 71 recordings from 3 operating conditions that were never used in training, and Macro-F1 was 0.943. On the same test, though, a baseline that never looks at current and uses only the recording time scored 0.986. So this score alone does not let us say the model can diagnose faults on real factory motors. Below is what we did, what came out, and which parts should not be trusted yet.
Plain terms
Picture a student taking a test in the same classroom with the same teacher. The score is high, but if every test sheet came in a different pen color, you cannot tell whether the student solved the problems or memorized the pen colors. In MCC5, the recording time played the role of the pen color. Whether this model can solve the problems in a different classroom, that is, on a different physical motor, has not been tested.
Why look at motor current
Depending on fault state and operating conditions, a motor’s current can change in magnitude, frequency content, and the relationship between phases. MCSA (Motor Current Signature Analysis) looks for clues of a state change in these electrical signals. Current sensors are often already installed on equipment, so it is worth checking how much diagnostic information existing measurements can give.
A change in current does not mean a fault, though. Load, rotating speed, supply voltage, and the measuring equipment also change the signal. A fault classifier first needs data structured so that these factors can be separated from faults.
What we trained
The data is MCC5-THU. The recordings come from a single 2.2kW three-phase induction motor test rig and cover 24 states: 1 normal and 23 single or compound faults. The original is published on Mendeley Data under CC BY 4.0.
The model input is not the raw waveform. Current recorded at 12.8kHz is cut into non-overlapping windows of 65,536 samples (5.12 seconds). For each window and each phase we compute a Welch power spectrum and take the log of 512 bins over 0 to 6,387.5Hz, giving a [3, 512] input. The MantisV2 encoder processes each phase, the three outputs are concatenated, and the model computes probabilities for the 24 states. A recording’s prediction is the state with the highest mean probability across its windows.
flowchart TB
A["Three-phase current recording<br/>phases a, b, c at 12.8kHz"] --> B["Cut into 65,536-sample (5.12 s) windows<br/>no overlap"]
B --> C["Per-phase Welch power spectrum<br/>0 to 6,387.5Hz, 512 bins, log"]
C --> D["One window = [3, 512] input"]
D --> E["MantisV2 encoder<br/>per phase"]
E --> F["Concatenate three phases<br/>probabilities for 24 states"]
F --> G["Average probabilities over all windows"]
G --> H["Highest state = recording prediction"]
During development we ran leave-one-condition-out cross-validation (LOCO), holding out each of the 9 operating conditions in turn. At this stage we compared a physics-feature classifier, frozen pretrained MantisV2 representations, and full MantisV2 fine-tuning. The released checkpoint is a single model retrained on all 208 development recordings.
Results
| Evaluation | Macro-F1 | What it means |
|---|---|---|
| MantisV2 fine-tuned, 9-fold LOCO cross-validation | 0.945 (worst fold 0.884) | Average of models retrained per fold. Not the score of the released file |
| Released checkpoint, 71 sealed TEST recordings | 0.943, 95% CI [0.885, 0.990] | Sealed before training and opened once |
| Physics features + logistic regression, same TEST | 0.923, 95% CI [0.849, 0.980] | Conventional baseline |
| Recording time only, same TEST | 0.986 | Confound-warning baseline that never sees current |
| Recording date only, same TEST | 0.127 | Date alone gets almost nothing right |
Only the four results from the same sealed TEST are plotted. LOCO 0.945 is a different evaluation and is left out.
On the sealed TEST, the released checkpoint got 67 of 71 recordings right. For 20 of the 24 states, precision and recall were both 1.00. Of the 4 errors, 3 swapped winding fault severity H and L, and the remaining 1 called an outer-race bearing fault a bent shaft. Winding H has F1 0.40 and winding L 0.57.
These numbers come from a small sample. Each state has only 3 TEST recordings, and voltage unbalance L has 2. A single error moves that state’s recall by 0.33.
Macro-F1 was 0.020 higher than the physics-feature logistic regression, but the 95% confidence interval of the difference is [−0.051, +0.094] and includes 0. It falls short of the pre-registered superiority rule (at least +5pp with a CI lower bound above 0), so we do not claim that MantisV2 beats hand-built spectral features on this rig. On the same TEST, expected calibration error (ECE, 10 bins) was 0.025, lower than the logistic regression’s 0.142. The model’s probabilities are closer to its actual accuracy. That advantage, too, was observed only on the same test rig.
Why the recording-time baseline scores higher
In MCC5, state labels and recording times are tightly coupled. 21 of the 24 labels were recorded on essentially a single day. Recordings of the same label are clustered in time, so knowing the recording time is almost enough to guess the label. That baseline scores 0.986. A baseline using only the date scores 0.127. The problem is not the date itself but the order of recordings within a day.
MantisV2 never received the recording time as input. Still, differences in rig state or measurement environment that changed with recording time may have left traces in the current. With this data alone, we cannot rule out that the model learned those differences instead of the faults.
Measuring normal and faulty states alternately on the same day under the same conditions, and reproducing the same fault on different physical motors, would have made a stronger test. This data could not prove that kind of field generalization.
Why we released it anyway
This release is a Research Preview, not a finished product ready for the field. The goal is to make it reproducible how we applied a pretrained time-series model to motor current, how we compared it with conventional physics features, and which data confounds can shake an evaluation.
The weights and code are released under Apache-2.0. The base model paris-noah/MantisV2 is also Apache-2.0. The model repository does not redistribute the original MCC5 data, so please get the data from the original provider.
The model card example looks like this. The input is three-phase a, b, c current in amperes, and the sampling rate must be exactly 12,800Hz. Resampling was not validated, so the loader rejects other rates.
import numpy as np
from thaki_mcsa import load_model, welch_views, predict_recording
net, labels, meta = load_model("mantisv2_mcc5_full_welch.pt", device="cpu") # verifies sha256 before loading
current = np.load("current.npy") # [3, n_samples], amperes, fs = 12,800 Hz
out = predict_recording(net, labels, welch_views(current, fs=12800))
print(out["label"], out["top3"], out["n_windows"])
Installation and command-line usage are on the Hugging Face model card. For reference, one 90-second recording took about 1 second on a general x86 server CPU. We did not measure on edge devices.
What real equipment would need
Before applying this to field equipment, at minimum check the following. Does the same fault repeat across different physical motors? Were normal and faulty states measured under comparable load, speed, and voltage? Do equipment IDs, sensor codes, or collection times stand in for the true label? Also check that the same waveform or recording session did not end up in both training and evaluation. Finally, measure false alarms and missed faults on equipment never used in development.
Scaling the model up does not solve these items. They have to be solved in data and evaluation design.
From model output to maintenance work
Even when the model flags an anomaly, that alone should not stop equipment or trigger maintenance automatically. One option is to serve a validated model through an inference operations layer such as Metis, and use Paxis to set up reviewer checks, work handoff, and approval steps.
This is a reference design. It does not mean this MCSA model was integrated with Paxis and validated in the field. The structure that ties sensor verdicts to maintenance work under the same principles is covered in Reading Equipment Health from Sensors, and Letting Paxis Open the Maintenance Work.
Summary
This model achieved a high classification score on MCC5. The same evaluation also showed that a recording-time baseline scores higher and that its gap over the physics-feature baseline is not certain. What motor fault diagnosis AI needs is not one high score but evidence that it separates fault signals on equipment it has never seen.
If you want to assess whether AI can work on your own equipment’s current or sensor data, contact ThakiCloud. We start by reviewing equipment types, sensor setup, and measurement conditions with you.
More industrial AI models in one place: ThakiCloud Industrial AI hub
Model: ThakiCloud/Thaki-MCSA-MantisV2-MCC5 (Apache-2.0, research preview). Base model: paris-noah/MantisV2 (Apache-2.0). Data source: MCC5-THU, Mendeley Data, CC BY 4.0. The Metis and Paxis connection in this post is a reference design, not a measured result of an integrated product.
