JoeDMM
Member
I am interested in creating/sharing ideas that minimizes downtime by monitoring process and device inputs to predict possible failures. I started with the Rockwell Plant PAX code and am trying to improve upon it.
Ideas:
1. A report that displays the top ten valves/motors counters to find the ones that might be cycling too often. A valve should not be cycling 10,000 times a week...
2. Monitoring heat exchanger's transfer coefficient for signs of build up and when it should be cleaned.
3. Monitoring motor's change in amperage over time for signs of wear.
4. Matching vfd speeds to pressure and flow meters for signs of leeks or pipe build up or flow meter issues.
5. Matching control valves to process values for drift and wear.
6. I have pumps that have cleaner injection to prevent build up on the vacuum pump blades, but too much cleaner is a problem as well. So I am trying to create code that knows when to inject and for how long for best affect. work in progress
7. A system pre-start check where each motor, valve activate, but only for the shortest time possible to prove a system is fully functional before main production startup.
8. A LOTOTO success report automatically generated.
9. A matrix of conditions that suggest the probable cause for failure for each device type. This is highly specific because not all devices have the same devices around it. ie. a valve open limit not active but flow detected.
10. I think alarm conditioning is very important to reduce the numbers of alarms when there is a failure and catching the first alarm.
11. Instantaneous system snap shots when there is an important alarm and keeping the last ten in an array with the date/time, status of input and output cards and all analog values. Wondering if it would be beneficial to capture the scan it happens and the scan after?
Ideas:
1. A report that displays the top ten valves/motors counters to find the ones that might be cycling too often. A valve should not be cycling 10,000 times a week...
2. Monitoring heat exchanger's transfer coefficient for signs of build up and when it should be cleaned.
3. Monitoring motor's change in amperage over time for signs of wear.
4. Matching vfd speeds to pressure and flow meters for signs of leeks or pipe build up or flow meter issues.
5. Matching control valves to process values for drift and wear.
6. I have pumps that have cleaner injection to prevent build up on the vacuum pump blades, but too much cleaner is a problem as well. So I am trying to create code that knows when to inject and for how long for best affect. work in progress
7. A system pre-start check where each motor, valve activate, but only for the shortest time possible to prove a system is fully functional before main production startup.
8. A LOTOTO success report automatically generated.
9. A matrix of conditions that suggest the probable cause for failure for each device type. This is highly specific because not all devices have the same devices around it. ie. a valve open limit not active but flow detected.
10. I think alarm conditioning is very important to reduce the numbers of alarms when there is a failure and catching the first alarm.
11. Instantaneous system snap shots when there is an important alarm and keeping the last ten in an array with the date/time, status of input and output cards and all analog values. Wondering if it would be beneficial to capture the scan it happens and the scan after?
