4 replies to this topic

[AJ1795]

Our process includes extrusion into a starch bed.  The starch is continuously recycled.  We add starch when necessary and record those lot numbers, but how does this fit into traceability and recall?  I think there are parallels here to emptying silos, but I can't come up with a rational solution to this problem!  Can we just balance the amount of starch the bed carries with the amount of time that it takes to use that amount of starch and assume that all of that starch is recycled during that period of time?  Currently, the dry process equipment is never wet cleaned so there is never a "break."  Should we be cleaning this?  I really don't want to introduce water.  In the event of a recall (knock on wood!), how do we prevent having to recall everything ever made?  Not sure it really matters, but our product is high sugar, very low water activity. 

[TomLovesStarch]

Hi

 

For cleaning / hygiene could you empty it annually and clean it using dry methods - brushes, vacuums etc?  This would be a chance to check the fabrication as well - that the seals are well fitting, impervious welds etc...

 

The cleaning could then be validated by including the starch in the environmental microbiology sampling programme.

 

For traceability, if you could calculate the capacity of the bed, the amount of starch used per batch / per kg of confectionery made, and then if you could calculate the turnover of starch, you would be able to trace a particular batch of starch to a limited number of batches of confectionery.

 

It really depends on the amount of "co-mingling" of the starch. It's tricky because normally, with a silo, gravity helps you out by pulling the older starch through to the bottom, which minimises co-mingling, but I'm assuming in the bed, there will be a lot of shaking and agitation of the starch...

 

Common practice with liquid silos is to consider the following batch in the silo is contaminated with residual liquid from the previous batch. Given the extent of the mixing of the starch, you could arbitrarily say the starch remains in the starch bed for another 3 batches (eg if the starch bed is holds 100kg, the starch from the initial fill will still be present after the bed has been topped up with another 300kg of starch...)

 

I guess the only way to validate that is to fill the bed with a different starch, and run the line as normal and test for the presence of that starch regularly, until it disappears. (say if you normally use tapioca starch, you could fill the bed with corn starch and take samples regularly and test for the presence of corn) - obviously that's not ideal unless you have a customer who would accept a change in the starch for a short period :s

 

 

Tom

[Snookie]

 to the forum Tom.   

[Charles.C]

Our process includes extrusion into a starch bed.  The starch is continuously recycled.  We add starch when necessary and record those lot numbers, but how does this fit into traceability and recall?  I think there are parallels here to emptying silos, but I can't come up with a rational solution to this problem!  Can we just balance the amount of starch the bed carries with the amount of time that it takes to use that amount of starch and assume that all of that starch is recycled during that period of time?  Currently, the dry process equipment is never wet cleaned so there is never a "break."  Should we be cleaning this?  I really don't want to introduce water.  In the event of a recall (knock on wood!), how do we prevent having to recall everything ever made?  Not sure it really matters, but our product is high sugar, very low water activity.

Dear AJ,

 

As you surmised, this traceability scenario has been queried on this forum before with respect to silos and other continuous/semi-continuous processes.

 

One approach discussed here elsewhere is to simply define lots by time / input codes, eg one lot per hour/day etc. This still requires some assumption(s) to be made as per previous post regarding residue of course (ie one lot back whatever). The degree of carry-over may be amenable to some kind of estimation perhaps?. You might also try a little searching here, eg silo.

 

Rgds / Charles.C

[freeromios]

Continuous recycling without ever fully removing the starch from your system (let's say during annual maintenance) is a very poor practice, both form hygiene and from traceability perspective. I strongly suggest that you should empty at least one a year. This will create an @annual@ batch for your starch processed products.

This will assist you with the food authorities in case you ever have a major recall in place.

Some assumptions like the ones mentioned above may be also invoked in order that you reduce the size of your batches.