Heavy Equipment Cycle Monitoring With Stage-Based Image Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Monitoring the operation of heavy equipment such as mining shovels and excavators is crucial due to issues like wear of components, detachment of teeth or shrouds causing damage, and the presence of large boulders that can harm comminution equipment or haul trucks, which existing methods fail to adequately address.
Innovation Solution
Implementing an activity classifier model using a neural network to analyze a sequence of images captured during the operating cycle of heavy equipment, allowing for analysis of the operating implement, load, and mine face, and generating activity labels to identify stages of the cycle, with subsequent aggregation of metrics over multiple cycles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional monitoring methods are used for heavy equipment, then the system is simple and easy to operate, but it cannot effectively detect component wear, detached teeth/shrouds, or large boulders
Solution Approach 1:
The patent replaces traditional mechanical monitoring systems with an optical-based image capture and neural network analysis system. Multiple cameras capture images of the operating implement, load, and mine face, which are then processed by a neural network to detect teeth/shroud wear, detached components, and boulders. This substitution of mechanical sensors with optical sensing and AI analysis enables precise detection of previously undetectable conditions.
Solution Approach 2:
The patent introduces an intermediary neural network processing system between the raw image data and the final detection results. The neural network acts as an intermediary that analyzes sequences of images, identifies patterns indicating wear or damage, and generates alerts. This intermediary layer enables the system to process complex visual information and extract meaningful detection data that would be impossible to obtain through direct mechanical sensing.
2Measurement precision
If no stage identification is implemented, then the processing is simpler, but the analysis cannot be performed at specific critical stages of the operating cycle
Solution Approach 1:
The patent segments the operating cycle into distinct stages (excavation, transport, dumping, return) and performs targeted analysis for each stage. The neural network identifies which stage is currently active and applies appropriate detection algorithms specific to that stage. For example, during excavation, the system focuses on detecting teeth and shroud wear, while during transport, it monitors for detached components. This segmentation enables precise stage-specific analysis without requiring complex omnidirectional monitoring at all times.
Solution Approach 2:
The patent implements dynamic stage identification that adapts the monitoring focus based on the current operating stage. The system dynamically adjusts what it monitors and how it analyzes images based on real-time detection of the operating stage. This dynamic approach allows the system to concentrate computational resources on relevant detections for each stage, improving precision while managing processing complexity through adaptive resource allocation.
3Reliability
If continuous monitoring of all operating cycles is performed, then more data is collected for better analysis, but the processing time and computational load increase
Solution Approach 1:
The patent applies partial monitoring by focusing computational resources on critical detection tasks during each operating stage rather than continuously analyzing all aspects of every cycle. The neural network performs selective analysis based on the current stage and detected anomalies, processing only the necessary portion of data required for reliable detection. This partial action approach maintains monitoring reliability for critical conditions while reducing overall processing time and computational load.
Solution Approach 2:
The patent implements rapid stage transitions in the monitoring process, quickly moving through different operating stages and focusing intensive analysis only when anomalies are detected. During normal operation, the system processes images at reduced intensity, then ramps up analysis when potential issues are identified. This skipping approach allows comprehensive monitoring coverage while minimizing processing time for routine conditions and concentrating resources when reliability is most critical.
Data Source
AI summary
A method and system are provided for monitoring operations of heavy equipment having an operating implement configured to excavate a load from a mine face. The method involves: capturing a plurality of images during an operating cycle of the heavy equipment; processing the plurality of images through an activity classifier model, the activity classifier model including a neural network having been configured and trained for generating an activity label based on a sequence of images in the plurality of images, the activity label associating the sequence of images with at least one stage of a plurality of stages making up the operating cycle; and performing at least one analysis using the sequence of images associated with the at least one stage, the at least one analysis including an analysis of the operating implement, an analysis of the load, and/or an analysis of the mine face.