Acceleration RMS Trend
Sensor Fleet
| Identification Name | Asset ID | Device Connection | Sensor Status | Model | Last Acquisition | Latest Reading | Severity | Recording | Admin |
|---|---|---|---|---|---|---|---|---|---|
| No sensors have sent data yet. Check that sensors are powered on and connected. | |||||||||
Sensor Condition Status
Acceleration Status
NORMAL
--
Velocity Status
NORMAL
--
Displacement Status
NORMAL
--
Overall Asset Health
NORMAL
--
Sensor Profile
- Sensor Type
- --
- Vendor
- --
- Model
- --
- Available Metrics
- --
- Frequency Source
- --
- Frequency Validation
- --
- Temperature
- --
- Last Update
- --
Acquisition diagnostics
--
Trend Stability Indicators
Engineering Trend Analysis
Velocity RMS Trend
Displacement RMS Trend
Sensor-calculated Frequency Trend
| Axis | Latest Frequency | Moving Average | Change | Trend Quality |
|---|---|---|---|---|
| X | — | — | — | — |
| Y | — | — | — | — |
| Z | — | — | — | — |
Visible Trend Summary
6-Month History
Hourly groups readings into time blocks and draws a box plot (median, quartiles, extremes). Minute and second resolutions draw every stored reading as its own point — no grouping, no quartiles. The server automatically sends a safe number of points for the visible period. Drag left/right to move through time or up/down to move values. Scroll over an axis to zoom it, or inside the plot to zoom around the pointer, or use the zoom buttons.
Box: lower to upper quartileMedianWhiskers end at the lower and upper extremesOutlier or single reading
Bearing Frequency Reference
Enter bearing geometry above to calculate BPFO, BPFI, BSF, and FTF reference frequencies.
Threshold Configuration
Sensor ID--
Sensor Type--
Location--
State--
Alarm limits apply to the selected sensor. Alarms compare the sensor's actual reading, not the smoothed line shown in the chart - a spike can trigger an alarm even if the chart looks smooth.
Record Sensor Data
1
SensorSelect a sensor from the Sensor Fleet
3
This changes only the CSV save interval. Live sensor monitoring remains fast.
- Active recordings
- 0 / 10
- Time recorded
- 0s
- Rows saved
- 0
- File size
- 0 B
- File
- --
Recording details and CSV help
- Live data connection
- --
- Save interval
- --
- Readings not saved between intervals
- 0
- Estimated 30-day file
- --
- All recordings: 30-day estimate
- --
- Delayed file writes
- 0
- Sensors being recorded
- None
- Saved on
- This computer — choose a location when recording starts
How to read the CSV
server_receive_time_utc_iso8601 is the time this Tremitus server received the reading.
sensor_timestamp_raw is the sensor's original timestamp. sensor_timestamp_meaning explains how to interpret it.
Units use spreadsheet-safe text: um, degC, and m/s^2.
System Logs
System Health
Live Vibration Pattern Recognizer
This screen compares the live vibration with examples you recorded and shows which pattern is closest right now.
What am I looking at?
The big number is the match score — how closely the live vibration fits the nearest recorded pattern. The label is the stable result, shown only after it holds steady for several readings in a row. Known pattern means it matches something you taught it; Unknown pattern means the vibration does not resemble anything recorded yet.
--
Awaiting prediction
How the result was chosen
What does this mean?
- Best match
- The recorded pattern your live vibration currently matches best, and how close that match is.
- Second-best match
- The runner-up match. If it's close behind the leader, the reading is ambiguous and the result can flicker between the two.
- Shown result
- The system's actual answer. It only changes after several readings in a row agree, so one noisy reading can't flip it.
- Recent checks
- Shows whether enough recent readings agree before the displayed result changes.
Awaiting pattern distances.
Recent results:
--
Recorded Patterns
What is a “pattern”?
A pattern is a named machine condition you want the system to recognise — for example Normal, Imbalance, or Bearing wear. Anything unlike the recorded examples is shown as unknown.
Select a sensor to manage labels.
Abnormal / unrecognised pattern is automatic - never add it as a label. Pattern recogniser training requires at least 1 label and 20 samples per label.
Record Pattern Example
How do I record a good example?
Run the machine in the condition you want to teach, choose its name, and press record. Save at least 20 examples per pattern.
Recorded Examples
Why does this matter?
Check that every pattern has enough examples. Learning cannot start until each pattern has at least 20.
No dataset yet.
Learn Patterns
What happens?
The system learns each recorded condition. Run this again after adding examples or changing the selected signals.
Status
Idle
Stage
--
Samples
--
Labels
--
Pattern Setup
Choose signals, name the machine conditions, record examples, then start learning.
Setup Flow
0 / 5 ready
Pattern Settings
Current Sensor / Device
--
Signals Used For Learning
Which signals should I pick?
These are the kinds of measurement the recogniser learns from. Velocity RMS is the usual all-round choice for rotating machines. Add more only if they help tell your patterns apart — fewer signals give a simpler, steadier model.
1 signal selected for learning
Fewer signals = simpler, steadier model. Changing signals clears old examples and needs retraining.
Matching
What do these settings do?
Sensitivity controls how close a reading must be to a recorded pattern. Confirm after controls how many readings must agree before the result changes. Check every controls how often the system checks.
Changing sensitivity requires learning again. Confirmation and timing apply immediately.
Settings version: -- | Last saved: -- | Sync: offline