Machine Learning on CANplus™ Engine Control Panels

In March of 2023, Cattron introduced Machine Learning capability on its CANplus™ CP1000 and CP750-E panels for engine and VFD control, delivering the next generation of industry-leading performance and reliability for equipment operators.

Machine Learning prevents costly repairs and downtime by automatically detecting patterns, analyzing and troubleshooting potential issues. Panels will learn the normal operating behavior and analyze patterns of the engines/VFDs, and pumps, automatically alerting the user when parameters are out of range. Machine Learning takes the guesswork out of setting warnings and fault levels and monitors defined operating parameters to ensure the pump maintains optimal performance.  

Explanation of Machine Learning for Pump Control

Look up machine learning online, and you’ll find a myriad of definitions and explanations ranging from highly technical to basic and everything in between. For pump control applications, machine learning enables your CANplus panel to learn and monitor your specific application rather than relying on the factory default settings of the engine or VFD. For the CP1000 and CP750-E, machine learning is a feature that helps the CANplus panel learn what Normal Operation is, so it can alert pump operators when the panel is outside the normal range.

How Machine Learning Works on a CANplus Control Panel

Once the CP1000 engine or VFD control panel and application are set up with the desired flows, pressures and settings, the CP1000 begins the machine-learning process. It cycles through the entire operating range of the application, assessing and learning about all sensors and engine or VFD data coming into the panel. The CP1000 uses the information it receives to set normal levels (green), indicating that things are going well and operating as expected. It typically takes three to five minutes to complete the sequence.

Once the machine learning process is complete, the panel will automatically start monitoring the system. It will identify when operational parameters trend outside the normal operating range, enabling preemptive action to address the problem.

It also examines fluctuations and deviations across all incoming signals to establish yellow warning and red fault areas. The panel will issue warnings and alerts when any monitored parameter exceeds the normal range for a specified period.

Machine learning on CANplus panels leverages powerful edge computing capabilities to monitor your application without telemetry. While not required, Machine Learning fully supports remote monitoring when telemetry is added. Each learned parameter has a custom SPN that can trigger notifications when used with RemoteIQ™.