Heavy Duty Machine Monitoring
Challenge:
Software Logistics worked with a company that recycles aggregates from torn up roads and parking lots. This company deploys large rock crushers that ground up material into smaller rocks that are used for many purposes.
The challenge they had was that when the rock crusher isn't running, their very expensive equipment is sitting idle. When it was running, they didn’t have a good way of determining how much material was being processed. These machines were also in remote locations where connectivity was a challenge and some of them were even mobile to process aggregate at the job site.
Solution:
-
Work with subject matter experts to better understand the entire process of handling the materials.
-
Understand the constraints of the problem in that connectivity was going to be an issue and our hardware and electronics needed to be in a harsh environment.
-
After gathering an understanding of the problem, we determined that we could monitor the load on the motors used as part of this process.
-
When those motors were running, we knew that the line was running and weren’t accumulating down time.
-
By monitoring how hard those motors were working, we could gather rough estimates as to the production levels at any one time.
-
This solution dictated an always on connection using cellular technology (NB-IoT). This technology is more expensive than WiFi since you pay by the data you send, but more reliable and easier to configure.
-
We monitored the load on the motor once a second and sent that to the cloud for processing.
-
After a base line was determined of no-load conditions, anything above that could be integrated or added up over time to determine the amount of work completed.
-
Once week reports were generated and emailed to plan supervisors as to the amount of downtime that was experienced and an approximation of the production.
-
To account for accurate down time, we needed to identify the working hours of the machines as well as any planned downtime, weekends and holidays.
Results:
-
Very accurate analysis of down time was identified and sent out via weekly reports
-
Based on time and tuning of the algorithm we were able to track within 5% of the best estimate of the manual process.
-
Using a sophisticated algorithm, we were able to sample work once a second and report data to the cloud using cellular technology for pennies a day.