OSIsoft: Explain when to use conversion factors. v3.2
0 Comments


The Tag Tot Function makes the assumption that all the totalization values being input are being measured in Units Per Day. Now, that’s a pretty big assumption and, of course, most process variables are not going to be measured in Units Per Day. Most Flows are not in Units Per Day. So, the Assumption of the PI Server is the same regardless. The Assumption of the PI Server is always going to be Units Per Day. Now, if the actual Engineering Units of the PI tag are Units Per Day you’re in luck. The Conversion Factor’s going to be very simple. There really is no need for Conversion Factor. Just a zer. . .err, just a 1. Now, the problem comes in when you have the Engineering Units are actually in Units Per Hour, or Per Minute, or Per Second, and, of course, PI’s Assumption does not change. It’s still in Units Per Day. So what you’re going to need is an, a Conversion that will convert from this Assumption of Day constantly to whatever the actual Engineering Units are. And it doesn’t matter whether you’re doing a 2-hour Totalization, a 12-hour, an 8-hour, or a 24-hour Totalization. It’s the same. Some people think well, what if it’s just a 12-hour total. Is it different? No. We’re not converting from an interval of time, we’re converting from the Assumption into Reality. And the Reality is if you’re assuming a Flow is measured in Units Per Day, but in fact it’s measured in Units Per Hour, you’re going to have to multiply that Value by 24. Let’s say you are a manufacturer of widgets, and you make widgets at the rate of 50. Well, what is that rate? Is that 50 per hour or 50 per day or what? Well, let’s say you assume it’s 50 per day, and you get a report that says well, today we made 50. Then you find out later on that wasn’t 50 per day, that was 50 per hour. Very simple. How many hours are there in a day? There’s 24. So, you simply multiply that number that you got by the number of hours in a day. So, let me show you another way of looking at it. Now, here’s an example in which we have a set of 8 minutes, in which we have had a Flow that has changed 3 times in 8 minutes. So for the first 3 minutes it was just the Flow was right around 3 Engineering Units, and then for 2 minutes it was 5 Engineering Units, and then for 3 more minutes it was 1 Engineering Unit. Now, if you knew right going into this that this is measured in Units Per Minute, if that’s what your Engineering Unit of the Flow is, it’s a pretty simple calculation. It’s this first calculation that you see here. You would just simp, simply take that 3 Gallons Per Minute, multiply that by 3, that would be 9 Gallons. You take the 5 Gallons Per Minute, multiply it by 2, ’cause see it was 5 Gallons Per Minute for 2 Minutes, so that’s now 10 Gallons + 9, that’s 19 Gallons, and then take this last number here, for 3 Minutes it was at 1 Gallon Per Minute, so it’s 19 Gallons Per Minute + 3 Gallons — total of 22 Gallons. Okay, so far so good. That’s a pretty simple calculation. You know, if you,if you know up front that you’re measuring Gallons Per Minute and, for example, you’re trying to figure out what came in during this 2-minute period, that’s real simple. For 2 Minutes it was 5 Gallons Per Minute. That’s 10 Gallons, right? Well what if you weren’t that well-informed when you started out this calculation. Instead of Gallons Per Minute, for some reason you thought that this was coming in in Units Per Day. Because, see, that’s exactly what PI’s doing. PI doesn’t know about Gallons Per Minute. PI is assuming Units Per Day. So, how does it do that calculation? It does it by looking at this 2 minutes here, you know, which the Value is 5, and it says oh, 5, that’s 5 Units Per Day, 5 Gallons Per Day. So, how much did I actually bring in in 2 minutes? Well, in 2 minutes, at the Rate of 5 Gallons Per Day, you know, it ends up being this 5 Gallons divided by, or excuse me, 5 Gallons Per Day times the 2 Minutes, but that’s at a Rate of 1 over 1440. It’s, you know, that’s one day’s worth of Flow divided by 1440 Minutes because we’re figuring that this is 2 Minutes, so 2 Minutes of Flow ends up being, you know, this 5 Gallons Per Minute divided by 1440 – a very small number. And we’re going to do that for all of these. You know, each one of these we’re basically taking the, whatever that Flow amounts to, we’re dividing that by 1440, because we’re assuming this is in Gallons Per Day instead of Gallons Per Minute. Now, to solve this problem, which you end of having to do, is to multiply the result that we get by 1440. That’s the number of minutes in a day. How did this all happen? Why, why is it that we’re making this Assumption of Units Per Day? It’s just that the Engineering Units Field in the PI Server originally started out as a pure Text Field. It still is for the most part, except in some of our other Applications like AF. So the Engineering Units is a simple Text Field. We assume all Flow Units are in Units Per Day for purposes of Totalization, even though it may say Gallons Per Minute. PI doesn’t really know that it is Gallons Per Minute, so it assumes everything is in Units Per Day. And if I can offer one more example. Imagine that you are making widgets. You’re manufacturing at a Rate of 50 widgets. Well, unless you know what the Rate is, 50 widgets doesn’t really mean anything. Now if you’re making 50 widgets per day, and at the end of 24 Hours you’ve got 50 widgets. However, if you discover then you’re making 50 widgets per hour, not 50 widgets per day, then you take that 50, you multiply it by the 24 hours you’ve been making widgets, and that’s how many widgets you have. That’s the Conversion Factor. And, by the way, it has nothing to do with the Interval of Time. If you notice, this Interval of Time, here, this is just an 8-minute interval. That doesn’t, the Conversion Factor we used, doesn’t matter whether this is 8 minutes or 8 hours or 8 days. We’ve going to use that same Conversion Factor. Essentially what it is, it’s a conversion between the Reality of a PI tag, Hours, Minutes, and Seconds, and the Assumption PI makes. PI makes the Assumption of Day for all of those. Now, thankfully, you’re only going to have about 3 Conversion Factors, or 4 Conversion Factors: these right here. Every once in a while you might see something weird like if you’re measuring a Flow Rate in Units Per Half-Hour, you’d end up having to make this 48 instead of 24. We see odd things like that, but generally we don’t. And one more thing, while you’re converting, for example from, let’s say you’re measuring thousands of Gallons Per Minute, and you want a report in Gallons Per Minute, not thousands of Gallons, you could just add some zeroes to the end of this and that would end up — just add 3 zeroes — that would convert both your, the proper Rate Conversion, and also convert from thousands of Gallons Per Minute into Gallons. One more consideration is what happens when we have data that’s missing. Let me draw a little Trend here. Let’s say, for example, you’ve got data that looks like this. You’re gathering data, everything’s going along fine, and then the data’s unavailable for a period of time, and then it comes back, and this is what your data looks like at the end of the day. Well, what do we do with this missing data? With the data missing here, we don’t just leave that blank. What we end up doing for this missing data for this period of time right here, is we take the average for the rest of the period. So we take the, this period of time, this period of time, we average that out, and that’s what we come up with to substitute for the missing data. So we, we essentially substitute the average of the rest of the period to produce a full Totalization for the given period of time. In some cases people say we don’t like that, we can’t report that to the state agencies, or the EPA, or the governing body. Well, we need to do is simply say that we don’t have the, the Totalization for that period. Well, one of the ways we suggest you do that is you can use this Percent Good Function. Okay? With the Percent Good Function, you can do an If-Then- Else Statement that says, you know, if the Percentage Good is not greater than, say, 98 percent of whatever the requirement is, then simply return Bad Value, so that that doesn’t go into a report when it can’t be verified.

Leave a Reply

Your email address will not be published. Required fields are marked *