PREDICTING EQUIPMENT FAILURES AT A BULK MATERIAL SHIPPING PORT

CLIENT OVERVIEW

Location: North Shore, Burrard Inlet, Port of Vancouver

Industry: Bulk Material Handling, Marine Transportation & Logistics

Company Size: 350+ Employees

Throughput: 24+ Million Tonnes of Goods Shipped Annually

Services: Reporting, Quality Assurance, Uptime, Asset Tracking, Alarms, Conveyor Monitoring

The bulk materials shipping terminal is located on the north shore of Burrard Inlet in the Port of Vancouver. It is a critical link in the supply chain that transports Canadian commodities to markets globally, with trillions of dollars of goods managed through the facility.

The terminal collects thousands of data points per second from a wide variety of site assets. Having an automated way to collect, present, and act on that data was essential to predicting downtime and scheduling maintenance to avoid unplanned operational interruptions.

THE CHALLENGE

The bulk material shipping port required a trusted dashboard and notification system to better manage downtime and operations interruptions. The terminal's previous alarming system was a static threshold-based system which, due to the noisy nature of vibration data, produced a large number of alarms resulting in alarm fatigue and difficulty extracting actionable information.

Common Challenges:

• Realtime Monitoring of Streaming Data
• Alarm Fatigue from Static Threshold-Based Systems
• Predicting Equipment Failures Before They Occur
• Scheduling Proactive Maintenance
• Managing Unplanned Downtime
• Extracting Actionable Insights from Noisy Data

THE SOLUTION

Using vibration data from motors, QCA Systems developed a machine learning based algorithm that could detect changes in the structural health of motors.

With the application of machine learning algorithms, the noisy data series is transformed into a new data stream that is only sensitive to changes in the underlying state of the motor. This model provides valuable information for the early detection of possible structural deterioration or defects, ultimately empowering the business to schedule maintenance better and respond to issues proactively.

Here's how QCA Systems supports customers in this area:

• Machine Learning Algorithm Development
• Vibration Data Analysis
• Realtime Monitoring and Visualization
• Predictive Maintenance Modeling
• Actionable Recommendations and Reporting
• ML Pipeline Integration
• Rockwell Automation Historian Integration

KEY RESULTS

• Realtime monitoring of streaming model for vibration anomalies
• Reduced unplanned downtime through proactive maintenance scheduling
• Machine learning replaced static threshold-based alarming system
• Alarm fatigue significantly reduced through intelligent data filtering
• Maintenance team can visualize realtime and predictive data
• Potential issues are identified faster and resolved proactively
• Maintenance planning is completed with greater accuracy