Process Data Retrieval Using OSI Software’s PI SystemA generous gift from OSI Software, Inc. allows us to archive process and laboratory data for use in other application software. The gift includes their Universal Data Server, interfaces to control system software, and a host of client software tools used to retrieve, analyze, and report process data. The value of this gift is approximately $257,000. In addition to this, technical support from OSI and Dow Corning Corporation was, and is, critical to the success of the installation. We are extremely grateful to these companies andthe individuals that made this a reality. I’m sure you will find the software very useful as well as convenient and easy to learn.Loading the PI Plug-in for EXCEL:In order to access the process data from the CM computer lab PC’s you will have to follow this short procedure to install the EXCEL plug-in. This procedure loads the PI plug-in to your NT user profile so it should be there next time you need to access data from PI.After opening Excel, from the “Tools” pull-down on the menu bar at the top of the screen, select “add-ins”. Then select “browse” and, in the p:\PIPC 1.9\EXCEL directory, select the file “pipc32.xll”. Make sure the PI DataLink box is checked in the list of add-ins and click on “OK.” You should have a PI menu item on the top menu bar in Excel, and this should be there every time you open Excel in the future.When you want to access your process data, click on the PI pull-down menu and select the “Connections” option. If “steam” shows up on the server list, select it and click “Connect”. If you are asked for a user account, the user name is “pidemo” andthe password box is left blank. If steam isn’t on the server list, click the “add” button and add it to the list.How PI data archive works:Data are filtered, compressed, and archived for long-term storage and retrieval. Because the values that are stored in the archive are filtered and compressed, the archive data may not exactly match the actual process data. While each point passes through the value stored in the archive, the exact time the point actually passed through the value may not correspond with the archived value. Likewise, any signal noise is filtered out so that only an actual signal is considered for transfer to the archive.Here’s a simplified overview of how PI archives a single value. Let’s say we are interested in the temperature of the reboiler. The ProcessLogix point, TI132.daca.pv was mapped to the PI point D1:TI132.PV during system configuration. PI was configured to poll the control system every 60 seconds to read the current temperature. The value received by PIfrom the control system is called a “Snapshot” value and is placed in a register until the next scan when it is overwritten by the next snapshot value. The old snapshot value is nowpassed on to the compression process. The compression process compares this value to a value currently stored in the compression register and, if it is significantly different, passes the previous register value to the archive with its time stamp. Now this new value (the onewe just got from the snapshot register) and its timestamp are stored in the compression register and, sixty seconds later, the next snapshot value is compared to the one we just stuck in the compression register. Let’s say it hasn’t changed significantly. This next snapshot value is discarded. The process continues and, eventually, a value is significantly MTU - Chemical Engineering DeptPSCC1/14/2019different from the one in the register. Now, the one in the register is sent to the archive while the new one is put into the compression register.Important to remember – Snapshot values are not compressed, so the value and timestamp should compare exactly with the ProcessLogix value and timestamp for the same point; Data retrieved from the PI archive are not ProcessLogix data and should notbe used for such things as loop tuning; PI only archives significant changes in values, then interpolates between archived data when returning values to your spreadsheet; PI must be tuned for each point and the “best guess” values I used for determining a significant change may need to be adjusted, but you need to tell me that.Using PI – Building a tag listEvery time you use the PI system, you must connect to the PI server. After connecting to the server you can import a list of tags from the PI system into EXCEL.Open EXCEL and use the “Connections” menu item in the PI pull-down to connect to “steam”. Once connected, from the PI menu select “Tag Search”. In thetag mask box, enter the tag names you wish to find according to the following formats:For Solvent Recovery - Distillation,D1:tagname.extensionFor PDMS reaction,R1:tagname.extensionFor Utility process variables,UT:tagname.extensionThe tagname noted above is exactly as it appears on the Rockwell Automation ProcessLogix system except the “.daca” or “.pida” is not used.Extensions you may be interested in are:.PV for process variable information;.SP for setpoint information;.OP for outputs;.MODE for controller modes.You can insert a ? for a single character wildcard or an * to call up anything havingthe first few characters you identify. As an example, suppose I wanted all the temperature related points for the distillation process. I would type D1:T* for the tag mask. Or, if I wanted all the PV values for the PDMS reactor I would enter R1:*.pv in the tag mask field.As an alternate means of searching, you can put a word that you think would be in the description of a particular point into the “Descriptor” box. For example, putting E-101 in the descriptor box would generate a set of points having E-101 intheir tag descriptor.After you have entered either a tag mask or a descriptor, click on Search. A list of points will appear in the “Search Results” dialog box. Highlight the points you are interested in. Hold the CTRL key while selecting to skip around in the list. When you have highlighted all you want, click on OK to transfer the list of tags to your MTU - Chemical Engineering DeptPSCC1/14/2019EXCEL spreadsheet. The list of tags is transferred to EXCEL starting with the location of the currently selected cell and working downward.To retrieve historical data:On the spreadsheet, starting with the tag list you just created, insert at least three lines above the tag list. In the top