Implementing predictive maintenance for robotic arms using anomaly detection

This use case showcases how Apex Global prevented unexpected downtime and improved overall equipment effectiveness by proactively detecting and addressing potential malfunctions in robotic arms. The company identified anomalies in key metrics such as joint angles, motor current, vibration, and temperature and how the maintenance teams can be alerted to potential issues before they lead to costly failures.

Customer success

Apex Global could reduce unplanned downtime by around 20% using predictive maintenance for robotic arms. The company could reduce downtime and increase production efficiency. It can also lower maintenance costs by performing preventive maintenance only when needed and improve safety by preventing catastrophic failures. The maintenance engineer can automate the remediations by configuring maintenance ticket creation to save time and manual intervention through workflow builder in BMC Helix Edge.

Scenario

Apex Global wants to implement a predictive maintenance strategy to identify potential issues before they cause failures. The company's production line relies heavily on robotic arms, and unexpected downtime due to robotic arm malfunctions can lead to significant production losses.

Workflow

Task	Role	Action	Reference
1	Maintenance Engineer	Set up the following: Name: RoboticArmAnomaly Description: Detects anomalies in robotic arm movements. Algorithm Type: AutoEncoder-Anomaly	Core ML model workflow
2	Maintenance Engineer	Configure the following ML model data: Select metrics: Joint Angle 1, Joint Angle 2, Motor Current, Vibration, and Temperature. Set the collection interval, for example, every 60 seconds. Add filters if needed. For example, by specific robotic arm ID.
3	Maintenance Engineer	Upload the ZIP file containing the pre-trained AutoEncoder model.
4	Maintenance Engineer	Optionally, configure alerts for different anomaly score thresholds. For example, set the thresholds as follows: Minor: score > 2 Major: score > 3 Critical: score > 4
5	Maintenance Engineer	Deploy the model to the relevant BMC Helix Edge nodes monitoring the robotic arms.

Results

The system monitors robotic arm data in real time and generates alerts when anomalies are detected, allowing maintenance teams to address potential issues proactively.