17 replies to this topic

[GMO]

I used to work for a confectionery company and I never got a satisfactory answer on how batch traceability was maintained in silos which were cleaned once a year (if that) and topped up rather than run empty (otherwise the plant would stop).

Anyone got any examples or help? I realise if it were bread for example, you'd probably accept that all product on the marketplace would be withdrawn with a week max shelf-life (but then what about consumers that freeze it?) But with confectionery it was up to an 18 month shelf-life. Do you recall everything?

Ok, a sugar issue is probably unlikely but anything is possible...

[herb b]

I would be interested in this topic also.

What we do now is assign another lot number everytime we receive a load of sugar, oil, etc. We are assuming that the silo contains several mixed lots in reduced percentages. At some point, the fraction remaining of an old lot is too small to be of interest. (I believe we roll off after 5 lots) Each receipt is around 50 to 60% of silo capacity so the mixed fraction is 40 to 50% of the silo volume.

For recall purposes, we could identify all product with a specific lot number from the date created forward. In the case of sugar, it could be a lot of product.

[monkeyman]

Our traceability system goes back as many steps as we require but by default we have it set at 5 batches in a silo (silo holds about 3 different lots at a time). The only real way I can see to control batches fully is to have 2 silos for the same material and alternately emptying each one.... not really practical for everyone I guess.

Once had a problem with a physical contaminant in bulk sugar and stopped and emptied / cleaned the silo but if this was something which could not be detected until later on then don’t know how it could be done.

[elie tekly]

do necessary analysis for the remaining goods and then consider them as part of the new lot

[nitac]

I have a similar problem with traceability of liquid in vats, constantly dripping in and then out as part of the process. One of our head office "boffins" for want of a better word did a 6 month research project resulting in a 100 page paper, mostly of calculations way above my head. Bottom line is that I need a 12 week recall window. However, this formulation or recall "batch" has never been challenged by auditors, think they are probably as confused by it as I am.

[Dr Ajay Shah]

This is an interesting topic from powder to liquids.

A lot of modern silos are computer programmed to indicate what has been loaded in them and from where it has originated and also information as to how much has been utilised. in addition one can also note when it was last cleaned looking at the cleaning and maintenance records.

I condust feedsafe audits in Australia and look at tracebility of these items very closely.

I have also seen a programme for liquids used in the winery industry called Wine File winery manager and this is good for tracebaility of liquids in vats etc.

http://www.winefile.com.au/

I hope the above helps others in their work etc.

Cheers

Ajay Shah

[GMO]

Thanks guys. The mixed batch idea for powders clearing after 5 deliveries (for a 3 delivery silo) is an interesting idea. Do any of you have any evidence or references to back this up?

[MRios]

I agree with Monkeyman. Having 2 silos for sugar would allow you to clean one of them when it emptied out, without running out of sugar. As for cleaning once a year, hmmm. It would have to be "dry" cleaning, in order to avoid microbial growth, but once a month would be more like it.
What about having a type of "holding tank" for batches that come in? This would allow you to run all tests before actually putting that batch into your common silo. I don´t know what type of sugar you´re talking about (granulated or fine), but if it´s fine sugar, you´re probably getting a funnel shaped hole in the middle of your silo, so that all batches of received sugar would mix.
As for recall, do you mean having to recall a product due to finding that the candy deteriorated during it´s shelf life? In 18 months, there will probably still be candy on the market, but a lot of it will probably have been eaten within 3 to 6 months, one would suppose, since clients can´t afford to have products in their warehouses or shelves during more than that, I suppose.
As for having issues with sugar, that would depend on your supplier. I have my serious reservations with wholesale sugar that is put into ships, in a way similar to grain. From what comes in grain, I feel just a bit uneasy with this method of storage / distribution. But it would take someone who works with sugar to deal with this point.

[GMO]

But having two silos isn't always an option. For chocolate in one plant I worked in for example, they had a 20 tonne silo for white chocolate and 25 tonne deliveries. (Make sure you're running lads when they get here! NO! DON'T STOP!!!! )

[Sherrie]

We tackled that issue by adding a start and end time and volume (or weight if either is available) to our batching sheets that use ingredients in either of our 2 co-mingled silos. This way we at least capture the time frame that it was used. We keep a log off all deliveries of that ingredient and note the date, time and approximate volume that was currently in the silo prior to delivery. In the event there is an issue we can pin point the effected delivery and cross reference with the times that the ingredient was used in a batch. Not perfect, but it's working for us so far.

[GMO]

But how do you account for mixing? Presumably there is some even in powders?

[Sherrie]

In the event of an issue you would have to recall everything made after the problem delivery. It's certainly not a perfect system, but our silos are not ever completely empty and the option of additional silos is, well, not an option. You account for mixing by assuming that every load ever put in the silo is in there, as long as the loads are good, everything is great. If a bad load comes in, the whole silo is now contaminated and everything made after the addition of the bad load is also bad. At this point, the silo obviously would have to be drained and decontaminated. What we count most on is the ease of communication in this day and age as well as the traceabliltiy requirements of our suppliers. If a load is contaminated we are going to know within 24 hours and will be able to capture all that product.

So I guess the short answer is, recall everything after the introduction of a bad load.

I like the theory that you can roll over batches after so many and I guess if you're dealing in powders that are added at the top and removed from the bottom that's logical. However, if you're dealing with liquids that are both added and drawn off from the bottom you can't really ever say with much certainty that any of it is ever completely gone. Sounds gross now that I really think about it.

[Sherrie]

I've been doing a bit of research, I'm writing a paper on this very subject and I have to tell you there really isn't much in the way of helpful information readily available. There are a couple of journal papers out there in the Journal of Food Engineering and the like but you have to buy them to read them online or go to a university library to read them. There is a short piece provided by AIB called "Strategies for Bulk Material Traceability and Recall" in the September/October 2007 AIB Update. It was the most reasonably informative I've found and they're not trying to sell anything. The suggestion was pretty much as suggested previously to keep track of what's going in and estimate what has gone out. The Food Safety Modernization Act also has provisions for the FDA to do some more research and come up with more traceability guidelines. Okay, I have overstayed my welcome in this forum, but as I said I'm writing a paper so it's a hot topic in my brain.

[GMO]

Thank you. Just with the BRC requirement for mass balance it seems difficult to accept estimates but if I have a decent reference to back it up, is this what you're talking about?

http://www.qualityas...ticle_id=101444

Hmm, might help.

[GMO]

Ah, I've found the PDF, this is better:


https://www.aibonlin...raceability.pdf

[Sherrie]

That's the one

I don't suppose BRC offers any guidance on the subject? We got SQF certification with our traceability system, maybe estimates will be okay for BRC.

[GMO]

Well I know plants who have achieved BRC with silos but I think it's been skirted over tbh. I just reckon it would take a different auditor though to ask the question a bit more closely...

[mgourley]

We do something similar. Call it the "Three Lot Rule" if you like. For both bulk liquids and flour:

If the most recent lot is determined to be contaminated, the previous two lots will be considered contaminated
If the previous lot is determined to be contaminated, the lot previous to that and the current lot will be considered contaminated
If the second previous lot is determined to be contaminated, the following lot and the current lot will be considered contaminated.

For bulk soy oil and HFCS this works out to roughly one month production, based upon how often these ingredients are delivered.
For flour, it's probably around two weeks.

All our products have shelf lives of 7-11 days, so it may be a bit of overkill but as mentioned above, communicat6ion is key. If a lot were determined to be contaminated and that information was communicated within 24 hours, obviously you would not have to include as much product.

Marshall