From Raw Data to Actionable Insights: A Day in the Life of Geotech Assist
In the world of geotechnical engineering, data is crucial. But what truly matters is how quickly and accurately that data can be transformed into actionable insights. Today, we're pulling back the curtain on Geotech Assist to show you exactly how our AI-powered tool processes raw observational data and generates valuable reports for mining professionals.
Richard Shellam
6/24/20244 min read
Note: The data used in this example is simulated but represents a credible scenario that could occur in an actual mining operation. We've created this realistic dataset to demonstrate Geotech Assist's capabilities.
The Raw Data: A Snapshot of April 12, 2024
Let's start with an example from April 12, 2024. On this day, our geotechnical team conducted a specific hazard inspection. Here's an excerpt from the raw observational data:
"12/04/2024: The geotechnical team conducted a specific hazard inspection for the hazard area on the 668mRL bench. During the inspection, they observed that the tension cracks identified earlier were widening, with lengths increasing to 4-5 meters and dilations reaching 15-20 mm. In some areas, the cracks were found to be joining across multiple benches, indicating a more extensive area of instability. Furthermore, a small rockfall was observed within the area, providing additional evidence of the deteriorating slope stability. Ravelled material was observed on the bench below
Based on these observations, the geotechnical team raised the risk level for the failure area to high. The monitoring data from the robotic total station supported this assessment, showing an average deformation rate of 3 mm/day across all prisms in this area, a notable increase from the previous inspection.
In response to the heightened risk, the geotechnical team implemented an immediate no-go zone extending 10 meters from the toe of the hazard area. This safety measure was put in place to restrict access and protect personnel from potential rockfalls or slope instability until further assessments and mitigation strategies could be developed. The geotechnical team emphasized the importance of adhering to the no-go zone and will increased monitoring and communication with the mining operations team to ensure the safety of all personnel working in the vicinity of the failure area. The hazard maps have been updated to show the no-go zone. The Mining operations shift team leaders were asked to discuss this at the toolbox talk.
Data Extraction and Structuring
Geotech Assist takes this raw text and extracts key information specified by the user, structuring it into the formate you a. Here's how the system interpreted and categorised the data:
Date: 2024-04-12
Descriptive Location: Hazard area on the 668mRL bench at the open pit mine
Coordinates: X: 203, Y: 82, Z: 668
Risk Rating: 4
Estimated Risk: High
Failure Mode: Raveling
Clean Up Instruction: Implement and enforce a 10-meter no-go zone from the toe of the hazard area. Update hazard maps accordingly.
Observation Summary: Widening tension cracks, a small rockfall, and ravelled material indicate increased slope instability.
Observation Type: Hazard Observation
Reported By: Geotechnical team
From Data to Insights: Geotech Assist in Action
Now, let's see how Geotech Assist uses this structured data to answer questions and generate reports. For this example, we've set the date range in our chatbot from April 1st to April 12th, 2024. Including the observation detailed above, there were a total of 6 observations input into the system during this period. Here are some actual interactions with the system:
User: "What were the total number of inspections completed?"
Geotech Assist: "The total number of inspections completed is 6."
User: "What are the total number of rockfall events?"
Geotech Assist: "The total number of rockfall events observed is 2. These events are detailed in the observations dated 2024-04-02 and 2024-04-12."
User: "Can you develop a report for me? I would like a brief 150-word summary followed by a table to show each of the recorded observations and their associated hazard rating."
Geotech Assist then generated a comprehensive report, which included this summary:
"Geotechnical Inspection Report - During the recent inspections conducted between April 1, 2024, and April 12, 2024, the geotechnical team observed various conditions in the open pit mine. Most inspections reported no significant issues, with deformation rates and other parameters within acceptable limits, resulting in low-risk ratings. However, two notable hazards were identified. On April 2, a minor rockfall was reported on the haul road, attributed to overloading. On April 12, a high-risk hazard was observed at the 668mRL bench, involving widening tension cracks and a small rockfall, indicating increased slope instability. Mitigating actions, such as implementing no-go zones and continuous monitoring, were recommended to manage these risks effectively."
Observations and Hazard Ratings:
The Power of AI-Assisted Analysis
Geotech Assist not only extracted and structured the data but also drew connections between different observations over time. For instance, it identified the progression of risk levels and the recurrence of rockfall events.
Moreover, the system demonstrated its ability to synthesize information from multiple observations to provide a comprehensive summary. This capability allows geotechnical engineers to quickly grasp the overall situation without having to manually review each individual report.
Room for Improvement: The Human Touch
While Geotech Assist excels at processing and summarising data, it's important to note that human expertise remains crucial. For example, in this case, the system categorised the failure mode as "Raveling" based on the observation of ravelled material. However, a geotechnical engineer might interpret this differently, recognising that the ravelling is likely a symptom of a larger potential failure rather than the primary failure mode itself.
This highlights an important aspect of using AI in geotechnical engineering: while it's an incredibly powerful tool for processing and analysing data, it should be used in conjunction with human expertise for the most accurate and comprehensive understanding of complex geological situations.
Conclusion: Empowering Geotechnical Engineers
Geotech Assist demonstrates how AI can significantly enhance the efficiency and effectiveness of geotechnical engineering processes. By automating data extraction, structuring, and initial analysis, it frees up valuable time for engineers to focus on what they do best: applying their expertise to interpret results, make critical decisions, and ensure the safety and stability of mining operations.
As we continue to refine and improve Geotech Assist, we're excited about the possibilities it opens up for the field of geotechnical engineering. By combining the power of AI with human expertise, we're not just processing data faster – we're paving the way for safer, more efficient, and more insightful mining operations.
Geotech Assist
From Inspections to Insights, Instantly
© 2024. All rights reserved.