Please, bear with me.
And don't get me wrong: I have no more knowledge than the rest of you regarding whether acrylamide in potato products is a true carcinogen. By "true carcinogen", I mean that, when eaten in reasonable quantities, it carries significantly higher risk of cancer than, say, traveling on a commercial airliner. Most of us do the latter without thinking much about cancer risk. And most of us are exposed to low level cancer risks all the time. To carry risk, it needs to be higher than this background "noise" of low risk.
I have drawn no conclusions as to whether acrylamide is cancer-causing. I just have observations and questions. Reference my blog entries from October 21 (http://potatoprocessing.key.net/2008/10/more-on-acrylamides.html) and August 4 (http://potatoprocessing.key.net/2008/08/acrylamide-unkowns.html) of last year.
I write this in view of a couple of recent articles: Canada is moving toward labeling acrylamide as a carcinogen: http://www.potatopro.com/Lists/News/DispForm.aspx?ID=2335&Source=http%3A%2F%2Fwww%2Epotatopro%2Ecom%2FNews%2Easpx; A new study from Poland that demonstrates that excessive chip/crisp consumption carries some measurably negative health effects: http://www.potatopro.com/Lists/News/DispForm.aspx?ID=2324&Source=http%3A%2F%2Fwww%2Epotatopro%2Ecom%2FNews%2Easpx%3FPaged%3DTRUE%26p%5FDate%3D20090219%252005%253a00%253a00%26p%5FRating%3D3%26p%5FID%3D2343%26View%3D%257b82D67804%252d78DD%252d499D%252dA872%252d014A50505992%257d%26PageFirstRow%3D26; and a Reuters report on the lack of a statistical link between acrylamide and breast cancer: http://www.foxnews.com/story/0,2933,500692,00.html
(Thanks to Paul at http://www.potatopro.com/default.aspx for his excellent news site. If you don't spend time on Paul's site regularly, you should)
So, again, help me out.
We have a study from a number of years ago from Sweden, that says mice can contract tumors when exposed to acrylamide in concentrations far above what would be expected in normal humans eating potato products normally. Or even excessively. We have another study from about a year ago, from the Netherlands, that demonstrates no link between acrylamide consumption and risk of gastro-intestinal cancer. And now a study that shows no link with breast cancer.
We have this recent study from Poland that says that eating potato chips (with acrylamide) excessively will increase the amount of acrylamide in the blood. This must be from their rocket science department! Also apparently from the same department comes the finding that eating far too many chips/crisps (with attendant fat and salt) is not good for the circulation. Hmmmm. Seems they have made a brilliant discovery of the obvious. Interesting to observe that the Polish study does not appear to validate any link between acrylamide and cancer.
And so we now have Canada marching to the same piper as California: Treat acrylamide in potato products as conclusively cancer-causing (like tobacco). Regulate it! Label it! Advertise against it! Pay no attention to the levels of acrylamide in that bran muffin- bran muffins are healthy! "Potato", however, is a word that should regularly be linked with "couch", and so is bad.
And so our culture expresses its prejudice and tendency to over-regulate.
I have not yet seen a study that actually statistically links the consumption of potato products by actual humans in relatively normal amounts with measurably increased risk of cancer. Until we see such a study, we should exercise some rational balance between potential (read: unknown) risk and known cost.
OK, I'll step down from the soapbox now.
Tim
Saturday, February 28, 2009
Acrylamide News
Friday, February 20, 2009
Bean Cutters, Swan Song!
"So, let's get real, Tim!", you might say.
"No one in their right mind applies bean cutters to 15 ton/hour finished lines, only smaller ones." So, let's take a look at smaller lines. Let's also look at those that produce 10mm cuts as a primary product, rather than shoestring.
One aspect of comparison is that ADR capacity is granular: It can handle about 12 tons/hour of product (10mm cut), wet basis at the ADR. So if your line is 6 tons/hour finished or less (finished capacity is roughly half of wet cut capacity), one ADR will handle your defect control needs. Put a sorter downstream of the ADR for the full ADR-First (tm) effect, but that is optional. The issue is: If you run lower finished capacity, the capital expense of ADR is the same as for the larger line. Eventually, you can get low enough capacity so that the bean cutter approach looks attractive. Let's see how low we need to go before bean cutters make sense economically:
The rough numbers (see my Feb. 7 posting below) is that ADR saves €33.833 per finished ton per hour per year in product loss, ignoring the value loss of shortened product from the bean cutter, based on all our previous assumptions. Let's assume a 3-year payback is attractive for most processors. Let's assume a bean cutter line segment (with sorter) costs €350.000 and an ADR-only segment costs €500.000.
Under those assumptions, you would need a finished capacity of less than 1.5 tons per hour for bean cutters to make sense. These days, I don't hear of anyone designing so small a line.
Also, then, it takes only 3 tons per hour of finished capacity to make two ADRs pay back. But you won't need them unless finished capacity is over 6 tons per hour. In the end, all this means is that if your line produces more than 1.5 tons per hour finished, and your economics align with our assumptions of the last posting, ADR is the right choice for defect control, rather than bean cutters.
To top it off, the ADR-only line can remove 23% more defects than the bean cutter line. So if your raw quality is marginal, ADR will keep you in grade when the bean cutter will not. That is not figured into the equation..... yet!
It seems that, in the end, bean cutters are best applied to... beans!
Tim
Saturday, February 7, 2009
Bean Cutters, Part Deux
So, let’s look at the numbers for bean cutters in a potato strip application (see last week’s entry), and draw some conclusions, OK?
My model is admittedly simplified. For example, it assumes that all strips that contain peeled surfaces substantially on their ends are 75mm long. And strips that contain peel surfaces substantially on their sides are 60mm long. Yes, I will make a more detailed model later, but for now this will do.
The results? When running through a sort-then-bean-cut-then-sort line, defect removal is 65.8% maximum, roughly the same as a sort-then ADR cut line (64%). Compare that with an ADR-only line (80%) and ADR-First (93%). If your need is product quality, and/or your incoming quality exceeds 20%, ADR-first or ADR-only is clearly what you need. The bean cutter? Only if your incoming defect is under 20% (everyone busts that barrier sometimes, these days) or your product quality requirements are not as high as your competitors, does bean cutting make any sense from a removal perspective.
Now let’s look at yield. The total cubes coming out of the bean cutter (including cubes put into the blancher) is over 62% white, less than 38% defect. Now, if you have a bean cutter system you might measure your white cube rate at “only” 50-55%. But you are not seeing the rest of your losses. Take a peek at what is going into your blancher, what you see may surprise you.
Now, the typical modern ADR outputs a cube stream of about 80% defect, 20% white cuts. Many folks achieve better, but let’s go with those numbers. And let’s focus on comparing a sort-then ADR cut vs. sort-then-bean-cut-then sort line; as we mentioned above, if your need is high defect removal, you have already made your decision.
The total flow rate of cubes out of the sort-then-ADR system is about 1/5th of the defect flow in. For example, if you are running a 15 metric ton per hour (finished) line, and the defect level averages 15%, your incoming flow of defect into the sorters is 4.5 tons per hour. And your cube stream is 900 kg per hour.
Now, the total cube generation of the sort-then-bean-cut-then sort line is 2.1 times that of the sort-then ADR system. So, for that same 15 ton/hour line at 15% defect, the cubes total 1890 kg per hour. The difference is 990 kg/hour.
Now you have enough to do the rest of the math. What is the value difference to you between cubes and strips? What is your flow rate? What is the nominal defect level you wish to target? My bet is that the installed cost of an ADR looks to have a handsome payback for you, even if the bean cutters would be free.
Let’s just use some rough numbers. Assume 8 months of storage crop, running at capacity 23 hours per day, running 13 of every 14 days. Total bean cutting operational hours are then 5126. Assume you don’t need defect control on fresh crop. If you run a 15 ton/hour line, your loss due to the difference between a bean cutter and an ADR is 5075 metric tons per year. Now, if the difference in value between cubes (going to formed or flake) and strips is "only" €0,10 per kg, the value loss due bean cutters is €507.500 per year. At 10% MARR simple annual compounding, the present value of that cost over 7 years (minimum life of equipment) is about €2.500.000!
That is enough to buy the ADR AND the yacht!
And this doesn’t yet address the other costs: you need more sorting capacity with a bean cutter line, due to the need to re-sort the cut strips, then resort again and again. And the other matter: With the bean cutter, your average strip length drops like a rock. You have many more short pieces in the line. You would probably struggle to make length grade for many products. You can do the math on that issue, as well.
That cheap bean cutter doesn’t look so cheap now, does it? Like I said in my last post, you can pay now or you can pay later.
So, we’re back to somber, serious, hard numbers after a blog entry that was perhaps a bit of fun last week. Which style do you prefer?
Tim
Sunday, February 1, 2009
Bean Cutters: The reality
Folks, sorry for the hiatus- You will definitely see me posting weekly going forward. Change you can believe in!
Let's have a chat about bean cutters. For those of you who don't know, a number of potato strip lines today run with bean cutters in lieu of ADRs. Bean cutters are devices that can be configured to cut a segment off of one end of long, thin objects (like snap beans, which is what the machines are made to cut). There is no selection of which end will be cut- it is a blind, random cutting of one end of the strips that enter the device. Typically, folks set them on strip lines to cut 15mm at a whack, but that is an adjustment that can be anything you wish.
They set their lines up with optical sorters, sending the rejects to one or more bean cutter. The remnants are then sorted again, the pass stream going to the blancher, and the rejects to the bean cutter again.
There is, of course, no mystery as to why people have done this. Let's be honest- ADRs are expensive and bean cutters are cheap. Easy math. No one wants to pay out big bucks if there is another way to get the job done. And, if you are an equipment supplier who does not offer ADR, what choice do you have? You sell what you can deliver! Tout its value! And who doesn't want low capital cost?
The problem is, just like the oil filter ads of a couple decades ago, "you can pay now or you can pay later". And the amount you "pay later" is, let's just say, eye-opening. Just like those filter ads of yesteryear.
I have spent the last couple of days modeling what a sort-then-bean-cut line does to potato strips. Without going into all the details, if you did to people what bean cutters do to potato strip yield, you would be in the slammer 20 years-to-life.
Just look at your average, pedestrian potato strip. 7mm cut. 75mm long. No criminal record. Defect on one end, like 15% of his relatives (5% have a defect in the middle, but that is jumping ahead). When he runs through the sorter, he is rejected and sent to the bean cutter. SNIP! He loses 15mm in length. Now, half the time, his defect end is cut off, and he passes into the blancher his second time through the sorter. Would that it were always so simple!
(Ever wonder which end is YOUR defect end?)
Problem is, half the time, the wrong end is cut. Whoops, there goes waste. A white cube down the drain! But wait, we're not done... the strip comes back around through the sorter, where 20% of the time he is missed and passed into the blancher! Holy cow! You mean, I could lose a white cut, and still pass the defect? Yep. Happens all the time. And even if the right end was cut the first time through the bean cutter, the good remnant can be rejected back to the cutter if it is lying too close to a defect going through the sorter. WOW! You mean, everything can go right, all through the cutter, but I can still end up making white cubes? You bet.
And we're STILL not done. If the wrong end is cut the first time, and the strip is successfully sorted to the cutter the second time, what are the chances the wrong end is cut again? 50%! But the strip is now only 60mm long and getting shorter. Recovery of the remnant is getting less and less valuable. And if the wrong end is cut a second time, there is also a 20% chance that the defect passes through the sorter into the blancher. And on and on it goes, until we have a big pile of nubbins (including some actual defects), and compromised product quality.
I think you get my point.
Next time, I'll work through some of the actual numbers with you. As you can tell, I was having a bit of fun with this tonight. Next time, we'll get serious. Promise!
Tim