Spark events during longwall cutting can provide important information about the cutting process, ground conditions and potential operating risk. When a shearer drum cuts into hard inclusions, roof, floor or difficult strata, visible sparking can occur around the cutting zone.

For operators, engineers and site personnel, these events are not always easy to monitor consistently. Visibility can be limited by dust, lighting conditions, camera position, remote operation and the harsh nature of the underground environment.

The JMB Automation Spark Detection System uses vision-based detection to identify spark events around the shearer drum in real time. The system can use either classical machine vision techniques or machine learning models, depending on the available camera hardware and the requirements of the site.

By detecting, timing, classifying and recording spark events, the system provides a practical tool for improving operational visibility, supporting steering decisions and assisting with safer, more consistent longwall cutting.

The Challenge: Spark Events Are Easy to Miss

Longwall shearer cutting is a dynamic process. Operators are required to make steering decisions while considering seam position, roof and floor conditions, cutting performance, machine position and production requirements.

In some conditions, spark activity can indicate that the shearer is interacting with material outside the desired coal seam, such as rock inclusions, roof or floor. Short spark events may simply indicate small inclusions within the coal face, while longer or repeated spark events may suggest that the shearer is cutting into harder material or drifting out of the preferred cutting horizon.

Without a dedicated detection system, this information can be difficult to capture consistently.

Common challenges include:

• Limited visibility around the cutting drum
• Dust and poor underground lighting
• Reliance on operator observation
• Inconsistent event reporting
• Difficulty reviewing what happened after an event
• Limited visibility during remote operation
• Difficulty identifying repeat spark zones along the face

The issue is not just detecting that a spark occurred. The real value comes from understanding the type, duration, location and pattern of spark events so the information can support better operational decisions.

The Solution: Vision-Based Spark Detection

The JMB Automation Spark Detection System uses live camera vision to monitor the shearer drum and identify spark activity during cutting.

The system can work with any suitable camera arrangement, provided the camera has visibility of the shearer drum and cutting zone. This allows the system to be applied across different longwall installations, camera types and site layouts.

Depending on the available hardware and site conditions, the detection method can be configured using either:

• Classical machine vision, using colour, brightness and image processing logic to detect spark activity
• Machine learning detection, using a trained model to identify spark events from camera images

This flexibility allows the system to be matched to the real operating environment rather than forcing every site into one detection method.

Once a spark event is detected, the system can assess the event duration, severity and behaviour. This helps distinguish between short spark activity caused by small rock inclusions and more significant spark patterns that may indicate the shearer is cutting into the roof, floor or hard strata.

How the System Works

The system continuously monitors the shearer cutting drum through a live camera feed. Spark activity is detected in real time and classified based on the event characteristics.

Spark events can be separated into different categories, including:

• Single spark events
• Repeated sparking
• High-severity events
• Long-duration events
• General spark detected / no spark detected status

The timing of each spark event is an important part of the system. A brief spark may indicate a small inclusion in the coal face, while longer-duration or repeated events can provide stronger evidence that the shearer may be cutting outside the ideal seam horizon.

This makes the system useful not only for spark visibility, but also for assisting with steering corrections through the coal seam.

Dashboard Visibility and Event Review

The system includes a live dashboard that provides clear visibility of spark detection activity and system status.

The dashboard can display:

• Live camera feed
• Spark detection overlay
• Spark count
• Event history
• Camera health
• Software health
• Detection confidence
• Shearer position at the time of the event
• Event timestamps
• PLC-linked status values
• Current detection state

Detected spark events are also recorded as video clips. This allows operators, engineers and site personnel to review spark activity after the event and better understand the cutting conditions at the time.

This is especially useful for incident review, operational improvement, remote support and identifying repeated problem areas along the face.

Heat Map and Cutting Insight

One of the key features of the JMB Automation Spark Detection System is the ability to identify repeat spark zones.

By linking spark events to shearer position, the system can generate a heat map showing where spark activity is occurring along the longwall face. This allows the site to identify areas where sparking is more frequent, more severe or more consistent.

This information can support:

• Steering correction decisions
• Review of cutting horizon control
• Identification of hard inclusions or difficult strata
• Better understanding of coal seam conditions
• Remote operation support
• Operational review and reporting

Rather than treating spark events as isolated observations, the system turns them into useful operational data.

Integration with Site Systems

The JMB Automation Spark Detection System is designed to integrate with existing site control and monitoring systems.

System outputs can be made available to:

• PLC systems
• SCADA or Citect systems
• HMI displays
• Site historians
• Databases
• Dashboards
• Alarm and reporting systems
• Shearer onboard PCs

The system can output spark detected status, spark count, spark alarm level, event timestamps, detection confidence, camera health, software health and shearer position at the time of detection.

JMB Automation can also provide direct communication with the shearer onboard PC where required.

The goal is to make spark detection information available where operators and control systems can actually use it.

Supporting Safer and Smarter Longwall Operation

The JMB Automation Spark Detection System is not positioned as a replacement for site procedures, operator judgement or existing safety systems. Instead, it provides additional visibility and decision support.

The system supports earlier visibility and response to spark events by giving operators and site teams a clearer understanding of when and where spark activity is occurring.

Key benefits include:

• Earlier detection of spark activity
• Better operator visibility around the cutting process
• Reduced reliance on manual observation
• Recorded video evidence for event review
• Spark event history and reporting
• Identification of recurring spark locations
• Support for TARPs and site procedures
• Improved understanding of cutting conditions
• Assistance with steering corrections through the coal seam
• Support for remote operation of the shearer
• Integration with existing PLC, SCADA and dashboard systems

For remote operation, the system is especially valuable because it provides another layer of feedback to help operators understand what is happening at the cutting drum without relying only on direct visual observation.

Continuous Miner Applications

While the system is primarily focused on longwall shearer cutting, the same vision-based detection approach can also be applied to continuous miner applications.

Where a suitable camera can view the cutting zone, the system can be adapted to detect spark events, record event footage and provide detection outputs for review, reporting or integration into site systems.

This makes the platform flexible across underground cutting applications where spark visibility and cutting condition feedback are important.

Standalone Product with Broader Integration

The JMB Automation Spark Detection System is a standalone product, but it can also be integrated with other JMB Automation systems.

This allows spark detection to become part of a broader operational visibility platform, combining machine vision, industrial control system integration, dashboards, event logging and site-specific reporting.

The system is designed for practical underground use, where harsh operating conditions, dirty cameras, vibration, dust and limited visibility are part of the normal environment.

Turning Spark Events Into Useful Operational Data

The value of the JMB Automation Spark Detection System is not just that it detects sparks.

The value is that it captures spark events, classifies their behaviour, links them to machine context and turns them into information that can support operators, engineers and site decision-making.

By combining vision detection, event recording, dashboard visibility and control system integration, the system provides a practical way to improve visibility of longwall cutting conditions and support smarter operation.

Delivering Intelligent Automation

JMB Automation develops practical automation systems that combine machine vision, AI and industrial control system integration for mining and heavy industrial environments.

The JMB Automation Spark Detection System is another example of how visual data can be turned into useful operational feedback, helping sites improve visibility, support safer operation and make better decisions during underground cutting.