Ace Up My Sleeve-Counting the Seconds and the Workings` of a Chaotic Mind

Can saving seconds add up? Emerson’s Anand Iyer provides his unique perspectives on batch processes and accumulated timesavings.

We were in a car, driving after work, discussing different things, that two engineers (one from Sales and one from Technology) can discuss after a late evening at a clients site. There had been a bus drive of 9 hours, just freshen up and breakfast and then the regular rigmarole… Gatepass… quick safety induction and finally doing control system technical work…

Anyway… we end up with this quip, “If we guarantee to save 1 minute a batch then the client says that they would give us the upgrade order.”

I never knew that sales were that simple… is it really that simple????

Say, the process manufacturer was to move from DeltaV MD+ to SX+ and I was able to set the execution time of the phases to 500ms instead of 1 sec. Would I have the batch time halved? No, probably not. The valves would just take the same amount of time to operate and process would be just as slow with the SX+ as with the MD+ … I’d better focus on getting good sleep…

But WAIT what would we save….We would save roughly 500ms per step and many actions… where there is no interaction but parameters are set or in the first passes and so on… That could add to a few seconds in a phase… And, with all that batch action and many phases, we could save many more seconds.

Some of them phases run in parallel… So we would have to take that into account… Slump.. Slump… Maybe this is not that good an idea…Better focus on a good night’s sleep….

So calculate the time savings per phase… see what phases are in parallel … I get a premonition that if I sit with calculations… Murphy’s Law would come into play and the initial parts (as with most substantial savings) would happen together and maybe some of the finer seconds would get lost as one phase or the other in that parallel action would be doing some field action….

Do I have a case or not??? I don’t know… for this I have to analyze the entire operation…..Maybe this is nothing and it’s better to focus on sleeping…..

OK! Just take the phase with minimum amount of time saved into account for parallel phases…

There is also the potential savings in field actions. This is questionable, as there is a 500ms savings potential of recognizing that the valve has closed. So per confirmation, there is a potential savings of 500ms…

So calculate the time savings per phase, take the phase with minimum savings for parallel phases into account, add the time savings in series… for field confirmations, calculate the savings at 500ms per confirmation and roughly take 50% into account ( a probability of 50%)…

Ok then we come at a theoretical timesavings figure….

So what could be my smart question if I am faced with such a scenario…

How much does a second mean to you?? That doesn’t sound all that good…And after several such questions… the best one seems to be..

How many more batches can you make if we save a minute in the entire cycle??

So if you are doing 12 batches a day, and you could save 60 seconds in a batch of 2 hours then you can make 30 more batches in a 300 working day cycle…

+300 X 24 X 60 X 60 / 7200 = 3600

+300 X 24 X 60 X 60 /(7200 -60) = 3630 after rounding the result…

Would 30 more batches a year make an “Upgrade” Sense….

As the bus rumbles on.. I seem to have an ace up my sleeve… A quick method of giving the first glance savings figure for moving from MD+ to SX+ for some batch processes.

  1. Per phase quickly look at the number of steps and non parallel actions, add them and then multiply by this magical figure of 500ms, We could do better if some of them have phases working at 2 seconds or more…
  2. For every field confirmation, add a potential 250ms…
  3. Take phase with lowest savings in case for parallel phases…
  4. Get a seconds savings figure…
  5. Take the batch time…
  6. Share with them how many “more” batches they could make…
Posted Wednesday, September 19th, 2012 under Process Optimization.

Leave a Reply