Rollup uses a Model Finder, which is designed to navigate existing twin hierarchies easily. Users can browse, review structural context, and select multiple models at once.

By default, Rollup combines anomalies across chosen models, groups them by property, and normalizes their scores into a single Anomaly Score.

Anomaly Rollup and TwinThread’s Platform are built to support enterprise-level scale. In one current customer environment, the Platform consolidates 80 active models across multiple different departments, with individual models connecting up to as many as 1,500 unique sensors.

Time-range alerts are featured in the Alerts Recommendations page.

TwinThread tracks anomaly duration using the “% in Anomaly” field. This calculates the exact percentage of time a property spent outside ideal operating limits within the selected timeframe. This helps engineers isolate chronic issues from noise.

Engineers can define a strict score threshold. Any anomaly failing to reach it is hidden from the display.

Yes, Rollup offers grouping controls that allow teams to segment data dynamically:

• Group by Digital Twin: Organizes the anomaly feed by individual twin, while preserving anomaly severity within each twin's list.
• Group by Topic: Organizes anomalies using operational topics mapped in the data, making it easier to review performance by process, department, or functional area.

Criticality is based on user-defined alert rules. Users set specific thresholds for up to four criticality levels: High, Medium, Low, and Info. When a property value or event state meets the configured threshold for a given level, the system automatically assesses and triggers an alert of that specific criticality.

The threshold buffer extends the acceptable range around alert thresholds, preventing alerts from being triggered too often. An alert only triggers once the condition has fully moved outside of the configured buffer (adjustable from 0-100%).

Cases and Notebooks (which can house SOPs) can be attached to alerts to guide actions. Users can set specific “User Recipients” or “Group Recipients” to receive email notifications when a certain alert criticality is triggered.

Note: An email notification is only triggered if the new alert has a higher priority than the last active alert.

Anomaly tracking on batch lines requires operational context on the exact manufacturing phase the asset is in at that moment. Because of this difference in structural data, batch anomalies must be analyzed within their individual model content.

Yes.