Baselines and Condition State

The baseline is a photograph of equipment's normal behavior. When compressor C-01 operates correctly, its temperature runs around 65°C and its vibration is at 3.2 mm/s. Those values are recorded as the reference. From that point on, every time the sensor sends a reading, the system compares it against that reference and determines whether the equipment is healthy, needs extra monitoring, or already requires attention.

What is it for?

Without a reference, it is impossible to know whether a value is normal or concerning. A temperature of 78°C may be perfectly normal for an industrial furnace but critical for an electric motor. The baseline gives context to every measurement for that specific piece of equipment.

Baselines allow you to:

• Detect gradual deterioration that fixed limits do not capture — the equipment worsens little by little without ever crossing an alarm threshold
• Evaluate condition in real time: good, acceptable, unsatisfactory, or unacceptable
• Calculate the condition sub-index that feeds the Asset Health Index (AHI)
• Identify when permanently changed operating conditions require relearning (replacements, retrofits)

How does it work?

The system analyzes the sensor's recent measurement history and calculates the statistical profile of normal behavior: the average value, how much it typically varies, and what the typical minimum and maximum values are.

With that profile built, it evaluates each new measurement by calculating how far it is from the normal average, measured in standard deviations. A 1-sigma deviation is normal variation. Two sigmas is a warning signal. More than 2 sigmas indicates a problematic condition requiring attention.

The system also recognizes the equipment's operating modes to avoid generating false alerts:

How to use it?

Learn an equipment baseline

The baseline should be learned when the equipment is operating normally, not during a failure or maintenance:

1. Go to the asset profile and open the Condition Monitoring section.
2. Select the event source (sensor) you want to learn the baseline from.
3. Click Learn Baseline.
4. The system processes available historical data and saves the profile.

Interpret the condition state

After learning the baseline, the system evaluates each metric in real time:

The system applies a worst-case principle when an asset has multiple sensors: if temperature is in good condition but vibration is unsatisfactory, the asset is reported as unsatisfactory. A problematic sensor is not hidden by the others.

View deviations by sensor

In the event source detail you can see for each monitored metric:

• The sensor's current value
• How much it deviates from the normal average (in absolute value and percentage)
• How many sigmas it is from the average
• The resulting condition

Practical example: The temperature sensor on compressor C-03 shows 73.5°C. The baseline indicates the normal average is 65.2°C with a typical variation of ±3.1°C. The system calculates that the deviation is 2.7 sigmas — unacceptable condition. The condition sub-index drops and the AHI falls automatically.

Relearn the baseline

There are situations where the previous baseline is no longer valid:

• After major maintenance: bearing replacement, motor rewinding, pump substitution
• Change in operating conditions: the equipment now works under higher load or at a different ambient temperature
• Replacement of a main component: the equipment's operational signature changes

To relearn, repeat the same learning process. The system maintains a history of previous baselines so you can compare how the equipment's behavior has evolved.

Key benefits

• Personalized reference per equipment — does not compare a gas compressor against an electric motor
• Detection of gradual deterioration before it crosses an alarm threshold
• Automatic real-time evaluation without manually configuring limits per metric
• The worst-case principle guarantees no problematic sensor stays hidden
• Baseline history to measure the impact of maintenance and modifications
• Weighted scores that feed directly into the AHI with objective data

Common use cases

Scenario 1: Detect bearing wear before failure Motor M-08 has been in service for 3 years. Its vibration baseline was learned 6 months ago: average 2.8 mm/s. Over the last 2 weeks vibration has progressively risen to 4.9 mm/s. The system reports unsatisfactory condition (1.8 sigma deviation) even though no alarm is configured for that value. The maintenance manager schedules a bearing inspection. The technician finds early wear and changes the bearings before the motor fails.

Scenario 2: Validate the result of maintenance Water pump B-12 had unacceptable condition (vibration at 3.2 sigma above average). The technician replaces the impeller and adjusts the mechanical seal. The next day the baseline is relearned with the equipment in good condition. The condition score rises from 0.10 to 0.92, which raises the AHI from 38 (Grade D) to 79 (Grade B). An objective record of the improvement is created.

Scenario 3: Avoid false alarms during startup Compressor C-02 takes 8 minutes to stabilize when starting up. During that time temperature rises rapidly and vibration is irregular. With startup mode active, the system widens the evaluation thresholds for those 8 minutes and does not generate condition alerts. Once stabilized, it returns to normal evaluation against the baseline.