Problem: What is a method that can be used to filter out a signal caused by tension for a motor output current trend?
Background: VFDs (Variable Frequency Drives) have become the default device for our clients to control motors. VFDs typically are equipped with the ability to monitor output voltage, current, power usage, etc. We believe that if we trend this data and analyze it, we can obtain useful information on what is going on with the conveyor belt the motor drives by comparing it to adjacent motors.
Details: We can create these trends and monitor what changes occur during various conditions on the conveyor belt such as the induction/discharge of packages, or when a jam occurs and packages begin to pile up. In addition we discovered that the current outputted from the motor without any load on the conveyor belt periodically goes up and down most likely due to how the belt was laced and tensioned. Attempts have been made to try to filter out affects on current due to tension but sampling a cycle with a period of the rotation of the belt with some success but it tends to deviate after some time.
Solution: We are looking for a method that can filter out the tension signal so the resulting trend becomes as consistent and as flat a line as possible when the conveyor belt has no load on it. This would allow us to more accurate understand what is going on. The method needs to be able to handle variations of the tension signal caused by environmental changes as motors don't run exactly the same in any given moment.
Note: In the Not Filtered Signal image you can see the periodic signal and the attempted anti-signal to get a result similar in the other image. The spike in the trend is due to the induction of a package followed by intentionally stopping the progress of the package simulating a jam.
About the recuiterMember since Jul 22, 2017 Ananya
from Maharashtra, India