5 of 5 people found this helpful
Being able to organise Assets and their relevant real time data in a logical structure is probably the instant benefit. Users not familiar with the tag naming convention of a facility but they know they want to find the discharge pressure of a particular pump they can easily find it from the hierarchical tree instead of searching tag names and more times than not, missing the tag description which validates their search. Not only real time data but also static metadata such as manufacturer, serial number, installation date, serial number, links to technical manuals you won't easily find without PI AF. Once you have added attributes like these you can easily search for all assets of a particular criteria. Imagine you had a notification from a particular vendor advising that a pump or transmitter had a known issue and you had this type of equipment spread out over multiple areas/sites or even continents? With PI AF you can easily search and find these as well as get their current operating states and parameters and conduct comparisons.
Once you have started your structure is when you put PI Analytics to work where for example you can setup a condition based monitoring system. You can then setup Event Frames linked to Notifications to advise when parameters are operating out of specification, maintenance scheduling or even to advise when production metrics have deviated from specification or plant trips you can receive instant emails or SMS text. As for Corporate Systems you can do things such as metallurgical accounting, inventory reporting, the list goes on. AF literally can be utilised and benefit any and every department within an organisation. As well as the groups you have mentioned, Process Technicians and Engineers are very big users of PI AF as they continually look for opportunities to improve efficiencies or to find root causes of equipment degradation or failures.
Once you have AF setup and working you'll wonder why you didn't do it a lot sooner.
A good place to start is to create a hierarchical template from a section of your area of interest. If it's IT, do a particular Server type and it's config followed by other relevant KPI's. If it's a production facility, template the end of the process area and work backwards or vice versa. My advise is always target the low hanging fruit first. If you find that the production team are not getting their reports out in time then focus on that need. You will then get some runs on the board and hopefully gain some advocates for the solution. Before you know it the areas of need will come flocking to you to implement similar applications. I have a background in Power Generation with many identical assets at a site such as gas turbines. For me is to template one Generating Unit then duplicate across the others via the template so analytics and comparisons can be conducted quickly and easily.
1 of 1 people found this helpful
I want to thank you for such a comprehensive response. I too work in power generation with three technologies, Wind, Hydro, and Thermal (gas turbines). I can't help but compare/contrast some of AF's functionality with our CMMS, particularly the equipment records and associated tables. CMMS's have, as a foundational attribute, the functional location/equipment hierarchies. Others prefer to call this the asset registry. All other CMMS activity springs forth from the hierarchy culminating with the Work Order. How we manage Work Orders is an entirely different subject and I could spend days talking about this.
The comparison I see is with the hierarchy. CMMS hierarchies loosely follow the P&ID's to include equipment some viewable via the DCS, sending a 4-20 ma signal to a node, while others, completely devoid of remote observation/control/alarm functions. Manually operated valves comes to mind (even those have limit switches). The other end of the spectrum contains instrumentation monitoring a process variable and subject to granular observation by several means. I envision PI AF aligning closely with instrumentation. I see where CMMS's limitation is the beginning of AF's features. What I ultimately hope to accomplish is to have AF send an alert that will trigger the creation of a Corrective Notification within our CMMS, the precursor to the Work Order, undergoing scrutiny before being elevated to a Work Order. Besides this, I would eventually like to inactivate as many calendar based PMs as well, and have meter-based/event-based PMs rule the roost in our work management system. I also want this to be home-grown product, much like our Asset Management Group - Failure Modes and Effects Criticality Analysis template we created.
2 of 2 people found this helpful
Sounds like you have a good vision to start with Edward. I too have come from Wind, Hydro and Thermal plants so I appreciate the journey you are about to embark on. You have every reason to be very excited!
Don't be afraid of including your assets that have no DCS feedback or integration such as hand valves, gauges, etc as they too will add value when combining that metadata with other adjacent assets or like asset comparisons. If you can get an export of your CMMS hierarchy you can bring this into AF and then start massaging and improving it from there, it will give you a good starting base. Also don't forget, one of the great things about AF is you can create multiple databases so you can have one for CMMS, one for metering, IT Infrastructure, etc. Make sure you stand up a test environment as soon a possible or when your production system starts becoming integral with daily operations. The sooner you can get your subject matter experts using AF the sooner you will see more value realized. Give them a sand pit to play in!
Depending on the vendor of CMMS you are using there are integrators from OSIsoft and others that allow transactions to execute work orders as you have indicated above. Excellent that you want to grow this internally as much as possible as it gives great ownership and strong support. As others have mentioned, go over the example kits and industry presentations to give you as much food for thought as possible. Also, engaging a PI System integrator with relevant industry experience will help you fast track and hopefully steer you away from making the mistakes others have made before you.
Excellent video, concise and a real good "big picture" from the 50k foot level.
5 of 5 people found this helpful
This is a great video to introduce you to the benefits of PI AF, the video is 4 minutes long but very informative: https://www.youtube.com/watch?v=VWrxS48PpjY
Now I will try and answer your other question on where you should start.
To answer this question, I will have to ask you a question. What is the business problem that you want to fix or monitor?
The reason we ask this, is because in the success cases we see that the customer has a business problem they wish to solve and this is the driver for the solution.
For example, one of your wind turbines requires more frequent maintenance than the others and you want to find out why. Now we have a business problem.
With this information we can start to think of what information from the turbine can we track and possibly compare against a perfectly working turbine or compare the gathered data about that troubled turbine to the norm of all turbines.
When you start to track specific information about an asset, as explained in the attached video you can create analysis and events that will help monitor the asset and all of its attributes.
There are courses that we teach and also workshops that we run using your business problems and your data. I think it would be very beneficial for you to contact your account manager and inquire about the workshops.
In addition with what we have been saying another way to help fast track AF development is through the use of AF development kit: AF development kits
The AF kits each have a specific business problem they are proposing to solve. They start with a video to show the problem and tell give you an example AF structure with analytics and attributes that are usually used to solve this problem.
From your profile I can see you are from the power generation industry and I have highlighted some posts that might interest you the most:
For your next question, yes corporate IT will play a large part of this. There are servers, domains, domain account users and domain service accounts that will have to be created. Also ports to be opened and downloads and installation. But bear in mind that IT doesn’t come in the initial phase as the main driver of the solution is the business case. IT has requirements such as security but they usually come when the solution is a little more developed.
I recently completed the "Building Asset Hierarchies with PI AF" online course so I am a newbie at this. I find this subject fascinating with many practical applications. I now know about the AF development kits and will explore these in the coming months. I see that these kits do not solely belong to generation, but T&D can also reap benefits from deploying PI AF. Thanks very much for introducing me to example kits.
If you would like further information or help, please don't hesitate to ask.
1 of 1 people found this helpful
Edward, first - see the other responses; they provide a lot of insight and food for thought.
So... I came from a utility environment and spent 2 years working in AF specifically on renewables: wind and solar.
Within PI, there were over 1400 wind turbines with multiple performance equations per turbines.
Looking at just one type of calculation, expected power, there were over 1400 individual PE tags; they were LARGELY similar, with slight differences.
When a change in the calculation for a site was required with 60 turbines, then 60 PE tags needed to be updated - time consuming and error prone.
Using AF, the calculations were put in templates; if a change was required - one change was made and applied immediately to all affected turbines.
For the primary input (wind speed), there was a primary input (turbine wind speed), secondary (wind speed from the site weather station) and tertiary input (wind speed from a nearby site with regression applied) - it was relatively easy to built the complex calculations in AF where it would have been a logistical nightmare to build and maintain.
Beyond that, event frames were used to capture 'batch' data when a fault was indicated at a particular turbine to capture lost production, etc.
If a site is under manufacturer warranty, and the fault is determined to be a manufacturer issue, the lost production can become a warranty loss and applied back to the manufacturer.
When a compliance issue is suspected, having detailed data can reduce or eliminate fines.