There were some questions remaining from the PI 101 talk during the Academic Symposium. I have attempted to answer those questions below.
PI on Raspberry PI?
Yes, while you cannot install the PI server on a Raspberry PI. You can use Raspberry PI’s to push to or receive data from PI
But AF? It allows you make a relational database of PI.
AF stores it’s configuration information and metadata in a relational database. The time series data is still stored in the PI Data Archive.
What about data quality?
When data quality is available from a data source, PI stores that data quality. It uses the data quality for doing a few things, including calculations involving PI time series data.
How would a K-12 PI curriculum look?
This could be done in many ways. San Leandro High School’s Physics Club did a project where they looked at energy being used at their district admin offices. Another interesting place could be a robotics club. We also see the IoT devices, such as the RoomBot to be a good way to introduce PI to K-12. This would be a great topic for a discussion feed.
If your trying to pull say ... MW data from PI system... what classes do you recommend attending at this conference to learn how?
Even though the Users Conference is no longer open, there are lots of ways to learn about getting data out of PI. If you want to pull the data into Excel or a display, we recommend “Visualizing PI System Data”. To do this programmatically or using PI Data Access Tools, we’d recommend “Programming in PI WebAPI”, an online course.
Is the PI system agnostic to the quality of sensor data. It does not pass judgement on whether data are reliable?
If the sensor is able to qualify the data quality, then PI can record it and take it into account in calculations.
What are the benefits of using PI server instead of a Big Data solution for real time analysis (HDFS, MapR, HBase, others)?
We are still researching the functionality of the other solutions and how they compare to the PI System. PI has the ability of gathering data from many data sources and can efficiently handle large amounts of data at a time. We will update this response as we investigate more.
Is data between polling intervals extrapolated? If yes, how is this validated in a regulated industry?
The PI System stores actual values from data sources and the times that they occurred. If the connection to the data source was stopped, it can show that as a break or gap. If you wanted to view data on a plot, it would extrapolate between the recorded points. If you want to pull into excel and see values at specific time points (and there isn’t a value in PI at that exact time) then it will extrapolate. Since PI doesn’t store extrapolated values and makes it very clear when displaying what is extrapolated an what is stored, it is tried, tested, and trued in validated/regulated industries.
Is PI to MATLAB free for the Academic Community?
The PI to Matlab utility is free to both the Academic Community and industrial customers. Here is the link to the blog on it, which includes instructions on use and downloading it (https://pisquare.osisoft.com/groups/academic/blog/2016/07/25/announcing-the-pi-to-matlab-utility). For other ways to get PI data into Matlab, see the white paper on PI data into Matlab (https://pisquare.osisoft.com/docs/DOC-1305).
Can we dump out the MATLAB outputs into PI archive or AF?
Yes, there are ways to do this. The White Paper noted in the previous answer helps that (https://pisquare.osisoft.com/docs/DOC-1305).
Can you push data to Python?
There is not an out of the box solution for this. You can pull PI data into Python, but to have data pushed to Python and trigger a calculation, a scheduled service would likely be used.
Isn't new version releases tied to User Conferences?
Our developers work around the clock to provide the most up to date software. We release on most products a few times a year. You can view the engineering plan. Sometimes they do make very large software releases around the time of the User Conference, but they aren’t held until that time period.