I would like to create a calculated tag that takes a moving average of another tag. For instance once I get a tag that counts how many breaks we have in a day, I would like to trend a moving average over a month or a year to see which way we are trending. How would I make that tag?
↧
Creating a calculated tag that takes a moving average of another tag
↧
Snapshot file (piarcmem.dat) size.
Hello Team,
In my current PI System the piarcmem.dat is almost hitting the 1GB file size. I tried to search for more information about how much is the recommend size for this file but I was unlucky. Our PI System has ~55K tags and I want to be sure that the file size of my system is accurate or I need to tweak something to enhance the robustness of the system.
Thanks in advance,
James
↧
↧
PIML Tours
I am in the process of setting up PIML tours for my plant.
Part of the plant can be run by one of two pumps. Is it possible to set the tour up so the operator can select which of the two pumps is running and then enter data for that particular pump, and then continue on the rest of the tour. There are only a few measurements to be taken before and after the pump, so setting up three separate tours is unfeasible and cumbersome.
Thanks
↧
Check UserID of the entity who has updated the PI tag
Hello,
Is there any way to check the user's ID who has updated a particular PI tag?
I am using PI-SDK and OLEDB queries to update PI tags in a .Net application, where I need this.
For example: If I am logged in to my windows account- "GuestUser" and I opened the application and updated any particular PI tag, is there any way to trace the user ID?
I tried using "Message Logs" in PISDK utility but not found anything such.
Thanks,
Souvik
↧
Where best to install EFGEN Interface
Hello: I'm trying to figure out where best to install the new EFGEN interface. Traditionally, we've used PiBaGen on our PI Server. Now, we have PI2012 and AF2012 (w/ SQL Server Std.) on one box with an interface node in our North building and an interface node in our South building. I'd consider this a "small" application with 4000 tags with snapshot and archive rates less than 4000 events/sec. I anticipate ~1000 elements and 12000 events/month. I've read you should put EFGEN on an interface node, but I get confused when we have two interface nodes. Should I install an EFGEN instance, per node? Can I put it on the PI Server node? What are the loading considerations?
Thanks!
↧
↧
Compression algorithm
Hi all,
I've tried to find the answer of my question in many places (this forum, vCampus, techsupport website, etc.) but couldn't find it.
I would like to know exactly what's going on when the PI System decided to replace the snapshot value when applying the compression algorithm.
What I understand is for a PI point on which compression is enabled, the System consider an archived value, the current snapshot value and is waiting for the next value (sent by an Interface).
When the system received this new value, it applies the compression algorithm.If this new value is within the acceptable range, then the current snapshot value is simply discarded and the new value becomes the snapshot value.
What I want to know exactly is : when is the "old" snapshot value discarded/deleted ?
I think that this value is discarded before the new value is considered as the new snapshot value.
I'm asking because I've configured a notification with PI Notifications that check the last archived value for a point and if the last event is 10 minutes old or more, it means that there is a problem with my interface. It happened so many times that my OPC Server wasn't communicating with the controller, but continued to respond to the PI OPC Interface requests. The only way I found to detect this problem is with the notification described above, using the controller's minute as the source tag. I've enabled compression to reduce archived value, so I'm expecting only 0's and 59's will be archived.
However, sometimes, I receive emails from this notification even I know that there are no problems with the OPC server or the Interface. I'm guessing it's only a question of timing (the time the old snapshot value is discarded, PI notifications check for the last archived value then the new snapshot value is updated.
What do you think about that ?
↧
Performance equation
Hello,
i am a new user of PI Performance equation
I am trying to create a tag with :
I need to concat a "result" and a string.
Ex . :
I have a tag "60Tankblend.pnt" --> result is "A308"
I need to create a tag : 60LEV"A308".PNT
Is it possible with Performane equation ?
Thanks
↧
Create a demo Point using SINUSOID
I am trying to create a new tag using point builder, and point this tag to the built-in SINUSOID to generate random data. I used Point Builder in PI System Management Tools. I created a new point with the following attributes:
Name: JV_Test_Tag
Point class: classic
Point Type: float32
Display digits: -5
Point Source: demo
Source tag: SINUSOID
The tag is not getting any data. The current value is "Pt Created", and the Archive Editor does not show any archived data.
I am very new to the PI system, i have worked with other Historians in the past. I have a vCampus subscription and have searched high and low for just basic information on creating points and cannot seem to find that information.
↧
Event Frame Searching Options in System Explorer?
Hello:
I'm just starting to get familiar with EventFrames. We have multiple extruders (Ext01, Ext02, ect.) at our facilty. I've setup my AF Elements with that in mind. I can get EFGEN to create my events. However, when I'm in SysExplorer I'm struggling to figure out how to quickly search by an element, or rather an Extruder. I was hoping, with minimal clicking to "find all events on Ext01, in the past 24hrs.". What am I missing?
Also, how do you delete several hundred events, quickly? I can delete one at a time, but not a bunch, at once.
Thx!
↧
↧
Totalizer - Counts offload resets based on another tag
First off thanks to any help you can provide. I need to create a Totalizer or some sort of tag that will increment one each time another tag goes active. Then it will reset when a different tag changes.
Details: Product offloads the machine and a sensor reads that the product passes. Gives a 1 then waits 1 second then goes back to 0. The totalizer would count that Tag. It would reset when the Tag that says product ID changes. This would give us the total count of Mattresses.
↧
PI to PI able to replace old value?
Hi All,
I need to sync sets of tags (~around 500 tags) in 2 PI Servers data and planning to use PItoPI interface.
It all goes well to push the data from PIServerB to PIServerA, but it is not updating the value if there is any changes on PIServerB.
There are several manual input tags from PIServerB, which the user might sometimes change the previous value.
I have read the manual, there is a tag configuration on Loc5=2, (Add event, replace if event a same time).
But the interface never picked the new value.
Below is my PItoPI batch file:
-----------------------------------------------------------------------------------------
.\PItoPI.exe ^
/host=PIServerA:5450 ^
/src_host=PIServerB:5450 ^
/ps=PItoPI ^
/id=1 ^
/sn ^
/f=00:00:05,00:00:02 ^
/rh=800 ^
/rh_inc=12 ^
/ec=2 ^
/ts
-----------------------------------------------------------------------------------------
And the tag configuration on receiving PI Server:
-----------------------------------------------------------------------------------------
Loc1=1
Loc2=0
Loc3=7
Loc4=1
Loc5=2
PtSource=PItoPI
Instrumenttag=TagA (source tag to be transferred)
-----------------------------------------------------------------------------------------
Am I missing any configurations?
I am using PItoPI ver. 3.8.4, PI Server vers. 3.4.375
Thanks * Regards,
Agus
↧
New Video on Pat Kennedy and the Real-time Data Infrastructure
Many of you already know about OSIsoft founder Pat Kennedy's focus on helping customers get value from real-time data. Here's a video to introduce those of you who are new to OSIsoft to Pat, and to his legendary focus on real-time data infrastructure.
If you're one of the folks who've been working with the PI System for a lot of our 30+ years, please consider replying below with any stories about how you found the PI System and what you've been able to do with it.
↧
How to plot 2 tags in one point (running data) in PI processbook
I want to show a point combining two tags (pressure and flow) of a Boiler Feedpump. Given the pump cavitation curve by the manufacturer, it can show by plotting a point using the pressure - (x axis) and flow (y-axis) of the running pump. In this way we can easily monitor if the pump is going beyond the cavitation curve limit to protect the from further damage.
↧
↧
Hours since tag's last value change
Good afternoon,
I am creating a PI display have has the current value for several dozen tags. Each of these tags may take on any value, and changes values at unknown intervals. Between changes the value remains exactly the same. I am trying to create a pi calc (or any other method that works) that shows the hours since that tag changed values (Any value to any OTHER value).
Is this a relatively simple problem? I tried extensively to search for a solution without success.
THANK YOU in advance
Example
Tag 1
Time Value
12:00 1.32
12:01 1.32
12:02 1.32
12:03 1.32
....
....
....
13:00 8.87
Results: value was the same from 1200 - 1300, so if you looked at this at 1400 (one hour after change) you would get "1" as the result in "hours since last change"
↧
Event Frame attribute calculation (avg. w/ filter)
Hello:
It takes us roughly 25 minutes to produce a part. Each part is setup as an event, via EFGEN. I can figure out how to add an attribute to an EF and show the average die pressure, for the event. HOWEVER, we need the ability to only include points into the average if the point is within an upper limit and lower limit. I can't figure out how to do an average while including a filter expression. Can you help?
↧
How to change archives size?
Hi,
I neet to change our "archive" file size.
Currently, we use 200mb for each archive
We would like to switch to 1gb. How and where I make the change?
Allan
↧
Time Based Trigger
How can I create a Totalizer tag that will count an offload sensor from the hours of 6AM-4PM and 4PM-2AM. I have the totalizer working I don't know how to write the final results based off the time given. I can create two tags if that is easier.
Essentially this is a shift counter.
↧
↧
Move PI AF element
Good afternoon everybody,
I would like to know if there is a quick way to move an AF element.
Example:
We have many plants and in each plant there used to be only one tank. In my AF architecture, I've created many AF elements of type "Plant" in which I've included a child-element of type "Tank".
Now in a single plant, we can have many tanks, so I've created a child-element to each plant called "Tanks" (from Template <None>). I would like to move the existing plant element from each plant into this new element "Tanks".
I don't want to create a shortcut (reference). The only way I have been able to do that is by deleting the existing tank and recreate it as a child-element of "Tanks".
↧
Datalink 2013 Tag Search...not very good
I am using Datalink 2013. The Tag Search window is no where near as good as the standard tag search window that is in every other PI tool that I have used (Processbook, previous versions of Datalink, SMT, etc.). Is there a way to get Datalink to show the standard PI Tagsearch form in DataLink 2013 instead of this "new" version?
Thanks.
↧
Advise vs Polled Tags
Background: PI Server 2012, OPC Interfaces installed on 5 nodes communicating with primarily Rockwell ControlLogix PLC's (a couple of Modbus ones in the mix is all). Using Matrikon OPC Server for AB for the Rockwell stuff and Matrikon OPC server for Modbus for the others. Currently about 1200 tags, but plan to expand the system to say upward of 5000 in the next 2 years.
Most of the tags that we setup to start with are Polled...with the majority of them set to 10 second scan class. I was nervous at the beginning that we didn't want to create to much traffic, events, etc. but have since realized that this is really not much of an issue on the PI server and doesn't seem to be an issue for the interface nodes either at this point.
Now that I know more about PI and how it works, it seems to me that Advise tags are really the way to go. I have a couple of questions related to this:
1. Can folks give me any feedback on why Advise is better than Polled (or vice versa) or really what the benefit of each might be in certain situations.
2. The majority of my PLC's are not doing anything over the network...other than talking with PI through the Matrikon OPC server (some are however)...but should I be concerned about loading on the PLC's if I were to simply set up most all the tags as Advise with 1 second scan class?
3. Anything I should consider or watch out for when I go back and change the scan classes on the current tags from Polled (10 second) to Advise (1 second)? I understand the "don't mix polled and advise tags in the same scan class and scan class 1 is reserved for advise tags).
Kind of looking for general expert guidance on this topic...how it works, benefits and drawbacks, etc. Any links to great articles or how to's would be appreciated too.
Thanks!
↧