So my immediate reaction to this is two fold:
1. What's the actual PE you're trying to evaluate? Does the syntax differ between the PE and the Analysis at all?
2. Confirm that attributes that you're pulling into the Analysis are the exact same tags that you're pulling into the PE. I've seen before wherein end users will use substitution parameters for tags and find that a template for an element is incorrectly setup which is affecting their analyses downstream, which is why I suggest double checking this.
In addition to Robert's suggestions, I would:
- If you are using a PI Collective, confirm that the data for the inputs to the PE is the same on both members.
- Confirm that the scheduling of the analysis and PE is the same.
Can you share which PE you are evaluating, the configuration and the version of Asset Analytics?
1 of 1 people found this helpful
Another possibility, if you are doing an aggregate over a duration less than five minutes, the analysis could be using all uncompressed data and the PE is using compressed data. To rule that out, you can preview the analysis results (archived data) and see if it matches historical PE evaluations.
You need to find out the exact data for the for the pi analysis of the results where you will be able to get the aggregate compressed data, where you will find out the performance of the equation which is going to have the windows 10 audio crackling, will guide you from which is going to be most helpful part where they will get the equation.