What would be the best strategy to start a PI IP implementation?
I'm happy to see that you brought this up - I'll be following this thread for sure. I've had this same question for a while now, and am in the middle of planning this as well.
First, list of actions and requirements of what we need AF to do for us. Second, we have a laundry list of server/tools/OS/PLC/gateways/etc updates needed and the snowball effect on the user tools that we currently use. Third, line up the right people at the plants/locations.
And between those steps, I had many calls with our OSIsoft account rep (Maxime McKay) in order to get things lined up, shared pre-reads and training links and to write the requirements properly. I think OSIsoft was excited to help us out with this.
What we ended up doing (yesterday)... (sorry for the OSIsoft plug - ha).... we're going with a prerequisite/requirement gathering Webex with OSIsoft and the AF Jumpstart program to train a specific plant's team and to fully setup an area of that plant with AF... tag hierarchies, templates, analytics, dashboards and the use of Element Relatives, BI Integrator, etc.
(I'm looking forward to reading everyone else's replies )
I think the strategy the Stephen Kwan always recommends is to start small and focus on one business use case. Find one very specific part of the plant with a specific problem and use AF to solve it. Start with templates from the start and Ales Soudek points out many times on PIsquare, just about anything ever created within AF should be a template. Also, work with someone who knows AF (or at least the basics) and include the subject matter expert on the plant assets.
I agree and build out everything for the small project (event frames, analytics, PI Vision screens, reports, etc.) so that everyone can get a feel for what CAN be done. Make sure that it is actually looking to solve a business problem and has good return on the investment/time you put into it. With that said, make sure your learnings have an eye to the future and you are asking yourself "how would we repeat this or scale this out?"
Rick Smith has done several great UC talks on this very subject. Go to the OSIsoft website and search for his presentations over the last two years.
Agree with Dan: find a champion project: good returns on investment, not too big, with people from the business that are driven to achieve results and share the lessons within the organisation
Retrieving data ...