Optimization of Quarry Operations and Maintenance Schedules
| dc.contributor.author | George, Brennan Kelly | en |
| dc.contributor.committeechair | Nojabaei, Bahareh | en |
| dc.contributor.committeemember | Barros Daza, Manuel Julian | en |
| dc.contributor.committeemember | Westman, Erik Christian | en |
| dc.contributor.department | Mining Engineering | en |
| dc.date.accessioned | 2023-06-29T08:01:01Z | en |
| dc.date.available | 2023-06-29T08:01:01Z | en |
| dc.date.issued | 2023-06-28 | en |
| dc.description.abstract | New technologies such as the Internet of Things are providing newer insights into the health, performance, and utilization of mining equipment through the collection of real-time data with sensors. In this study, data is utilized from multiple quarries and a surface coal mine collected through the software CAT Productivity and CAT MineStar Edge to analyze the performance of loaders and haul trucks. This data consists of performance metrics such as truck and loader cycle time, payload per loader bucket, total truck payload, truck plan distance, and loader dipper count. This study uses data analysis and machine learning techniques to analyze the performance of loaders and haul trucks in the mining operations used in the scope of this study. Data analysis of cycle time and payload show promising results such that there is an optimum cycle time for multiple loaders between 30-40 seconds that show a high average production. Furthermore, the distribution of production variables is analyzed across each set of loaders to compare the performance. The Caterpillar 992K machine in the rock quarries data set seemed to be the highest-yielding machine while the two Caterpillar 993K machines performed similarly in the surface coal mine data set. The Neural Network algorithm created a model that predicted the loader from the performance metrics with 90.26% accuracy using the CAT Productivity data set, while the Random Forest algorithm achieved a 79.82% accuracy using the CAT MineStar Edge data set. Furthermore, the use of preventative maintenance is investigated in the process of replacing Ground Engaging Tools on loader buckets to determine if maintenance was effective. Additionally, data analysis is applied to Ground Engagement Tools maintenance to identify key preventative maintenance schedules to minimize production impact from equipment downtime and unnecessary maintenance. Production efficiency is compared before and after maintenance on Ground Engaging Tools and concluded that there was no material change in the average production of the mine based on that analysis. The insights gained from this study can inform future research and decision-making and improve operational efficiency. | en |
| dc.description.abstractgeneral | New technologies are helping us better understand the performance of mining equipment. This is done by using special sensors to collect real-time data on information such as how long it takes for trucks and loaders to perform their job, how much weight in the material they can carry, and how far they have to travel. Through the use of data analysis techniques and machine learning models, the data are analyzed to investigate optimum performance metrics. An optimum time of around 30-40 seconds is discovered for the loaders to output their best performance. We also discovered that through a comparison of normal distributions, some machines in similar working conditions perform much better. In the case of this study, it was found that the Caterpillar 992K loader machine outperformed all the other machines. Using machine learning models, we could accurately predict the loader unit from its data with about 80-90% accuracy. Maintenance practices are analyzed on loader bucket parts that assist in digging to prevent unnecessary maintenance or loss of production. Through analysis of maintenance records and production, it was found that there were no big changes after maintenance was performed. This information can help fuel future research as well as show where improvements can be made. | en |
| dc.description.degree | Master of Science | en |
| dc.format.medium | ETD | en |
| dc.identifier.other | vt_gsexam:37849 | en |
| dc.identifier.uri | http://hdl.handle.net/10919/115571 | en |
| dc.language.iso | en | en |
| dc.publisher | Virginia Tech | en |
| dc.rights | In Copyright | en |
| dc.rights.uri | http://rightsstatements.org/vocab/InC/1.0/ | en |
| dc.subject | Mining Industry | en |
| dc.subject | Big Data | en |
| dc.subject | Optimization | en |
| dc.subject | Maintenance | en |
| dc.subject | Machine Learning | en |
| dc.title | Optimization of Quarry Operations and Maintenance Schedules | en |
| dc.type | Thesis | en |
| thesis.degree.discipline | Mining Engineering | en |
| thesis.degree.grantor | Virginia Polytechnic Institute and State University | en |
| thesis.degree.level | masters | en |
| thesis.degree.name | Master of Science | en |
Files
Original bundle
1 - 1 of 1