Mistake Proofing Solutions
- April 16, 2018 at 3:28 pm #55980
We make batch solutions which are sometimes diluted with IPA and sometimes with nPA. The batches are expensive and we had one diluted with IPA but it should have been nPA. The cost of this error is large enough to require mistake proofing, but we aren’t sure of a good method.
Does anyone have a good method to ensure that the correct solution from a chemical supplier is used? Currently we do barcode scans on each component and have a green light indicate that the correct items were selected, but the wrong label was placed on the incoming bottle. We have also had the solutions mixed first and then the barcodes scanned after only to find out then that the solution was made wrong.
Detection methods would be helpful if they are quick and robust, but we really should be making it so that this cannot happen.April 16, 2018 at 7:07 pm #202476
As you know, mistake proofing (poke yoke) involves designing the process so that it can only be done correctly, or at least that mistakes are immediately detected and corrected. From what you say, your problem is mistakes made by your supplier. It’d be easy to say you need better detection methods but what you really need to do is to work with your supplier and help them to prevent making such mistakes. This might involve assigning a team to help them, or paying someone to help them and sharing the expense. Getting it right is good for both of you. If they resist cooperating maybe it’s time to look for a different supplier.April 16, 2018 at 10:17 pm #202479
Thank you for reading and responding @Straydog. I should clarify that the chemical from the supplier was correct and correctly labeled. An incorrect dilution was made from it and subsequently labeled with our internal barcodes. This was an internal mistake. We buy the chemical off-the-shelf, so there is not internal barcodes when received. It was the first internal process which made the incorrect barcode. Is it possibly uncommon to purchase off-the-shelf when there are high stakes for a smaller company?April 17, 2018 at 8:18 pm #202485
Forgive me for making assumptions. So the product from the supplier is correct, and correctly labeled. You dilute it and relabel with an internal barcode which indicates the amount of dilution. The error is that dilution and internal barcode don’t match. So the question is how can you prevent that from happening. You might do an FMEA on this first internal process to understand how it can fail and where your internal controls are inadequate. Is there some way to automatically link measurements of the chemical and dilutant to generation of the barcode in order to eliminate human error?April 18, 2018 at 10:11 pm #202493
No need to apologize @Strayer. Thank you for revisiting the question. You have summarized it appropriately that the first dilution was incorrectly barcoded.
We can do verification after dilution and have the barcode print only if within specification.
I was hoping for suggestions on preventative measures before mixing. Imagine if there was a safety hazard if the wrong chemicals were mixed. Detection would be obvious, but it would be best to never mix them together in the first place. We are not a chemical company, so I was thinking there may be examples of how chemical companies avoid these errors.April 19, 2018 at 6:55 am #202494
There is instrumentation you COULD use if you’re adding the material into a process line. Other options also if not.April 19, 2018 at 8:36 pm #202495
@lausto Without knowing the details of your process it’s unlikely that any of us can suggest anything other than generalities. I stand by my previous recommendation that you do FMEA. If you aren’t familiar with Failure Modes and Effects Analysis you can find the template and guidance here and many other places via search. Basically you decompose the process into its elemental steps, identify the various ways each step could fail, evaluate each failure mode in terms of severity/probability/detectability to prioritize failure modes, then discover better controls to prevent or at least better detect failures. You’ll need to put together a team of people who are very familiar with the process commit to the time to do it. With the limited process you describe this might be a few hours or a day or two. I often recommend excluding managers, supervisors, and process engineers since they tend to be biased toward what they think is happening or is supposed to happen. You want people who know what actually happens. If you or no one in your company has done an FMEA you might do well to contract a consultant to facilitate the exercise. Since you’re dealing with a critical error this should be worth the expense. But beware of the consulting company game — Very often they’re more interested in getting a foot in the door and selling additional services so they only do enough to convince you that you need to pay for more.April 20, 2018 at 8:13 am #202496
There isn’t really much of a process to describe so I didn’t realize I was being vague.
We buy solvents in 2 L bottles from standard chemical suppliers. These are stored in a flammable cabinet. We dilute with water to different concentrations into 1 L bottles which are then labeled with a barcode and stored in a refrigerator for the next steps.
The operator has a checklist which outlines the procedure, but the error was grabbing the wrong solvent from the cabinet. That bottle was labeled correctly, but the operator simply made a mistake.
We have noted that
1. solvents have many naming conventions and the checklist had a different name than the bottle. We have corrected it to read the same.
2. The bottles all look alike. So we have added coloured tape to distinguish
3. The barcode scanning is relied on for most steps to prevent this, but the first step does not have a barcode. So this is our focus now – receiving will add a barcode. But I have argued that they could make the error as well.
4. This happened now as we have broadened our solvents. We didn’t have alternatives so we never encountered this.
All are good, but they still rely on a human doing the right thing. But maybe enough prompts add up to mistake proofed? I was taught that a train crossing with a sign is good, gates are better, but a bridge is mistake proofed. Is there a bridge option?April 20, 2018 at 10:49 am #202497
The FMEA suggestion has been fruitful. After sleeping on it, one engineer realized that the heat of dissociation is distinct enough that monitoring the temperature would detect immediately before we continue mixing an entire batch. We can interface the temperature monitor with barcode generation to remove human error. As long as they mix a small test first. But that is much more robust than our initial preventative measures.April 23, 2018 at 7:35 am #202504
It sounds like you want to implement a solution in raw material storage. There are a couple of things you could try to implement: Placement of raw material and/or varying packaging of raw materials. I suspect you do not have two flammable cabinets so placing the two material on different shelves that are clearly label for each could help. I would also suggest color-coding the raw materials and using the same colors to label in throughout the process could also help.
Sometimes companies name material with acronyms and going to extent of renaming two like acronyms where failure continues to happen can be work the trouble. Hope this information helps!April 26, 2018 at 6:01 am #202509
Ask yourself why are you diluting the alcohols. This adds to potential defects potentially unnecessarily.
You must be logged in to reply to this topic.