In the April 2008 Processing column, we discussed optimizing process lines by appropriate use of surge, as well as some other approaches. Continuing on that general theme, let’s consider optimum line speed and rework, inspired by some recent consulting experiences.
When Is a Continuous Process Not Continuous?
One measure of process performance is mean time between failure (MTBF) computed by taking the total time available for a run and dividing by the number of process interruptions. To be fair, a run can be defined as the time between deliberate stops, such as formula changeovers, weekends, or shift ends (if the next shift is not a continuation). In a recent example, MTBF for a supposedly continuous process was 10–15 min.
Mean time to repair (MTTR) is computed by taking total downtime and dividing by the number of interruptions. In this case, it was about 4 min, suggesting that most stops were due to jams and were easily fixed. However, stops in most lines are usually bad no matter how short they are. In this case, it was noted that products made after a line stop were usually off target weight and were discarded by a checkweigher. This phenomenon also contributed to relatively wide ranges in target weight, meaning that, on average, more product was given away than was desirable.
By law, packaged foods must meet or exceed label weight. Most fillers achieve a distribution in fill weights, depending on the consistency of the process and the precision of the filler. Companies normally determine the actual distribution, compute the standard deviation, assuming the distribution is normal, and set their equipment to target a weight that is about two standard deviations above label weight. Obviously, the larger the standard deviation, the more product is given away in the average package.
Reducing the standard deviation of filling particulate products is one reason for the popularity of digital scales, which can be very accurate and precise. Not all products lend themselves to digital scales, however. Some are filled volumetrically or by count, for instance. In this case, packages were filled by count, but labeled by weight, and after the line stopped, individual products tended to be light, so packages were light.
With such a low MTBF, it was a stretch to call this a continuous process. In fact, it was questionable whether the line ever experienced steady state. The first question was why was the line stopping?
By the way, the company in question kept pretty good statistics on downtime and causes—far from perfect, but probably more detailed than those recorded by other companies. One important lesson is the extreme importance of having good data if efforts to optimize a process are to be successful.
It turned out that most of the stops were due to a fairly complex step in packaging. Packaging of many consumer food products consists of primary, secondary, and tertiary operations. Primary packages are the ones that the food sees directly—cans, bottles, bags, canisters, pouches, and wrappers. Secondary packages enclose the primary and are not always seen by the consumer—cartons, cases, and trays, for instance. Finally, cases and trays are usually palletized and stretch wrapped to make a unit convenient for shipping.
--- PAGE BREAK ---
It is axiomatic that packaging should not stop a process line that is meant to be continuous. This means there should be carefully designed surge capacity or redundant packaging capacity. In practice, however, it is common that incremental line improvements erode whatever excess packaging rate there might have been and reduce the effectiveness of surge, unless it is increased at the same time. Often, space, time, and money constraints prevent this from happening.
A further complication is that packaging materials—especially those that are not seen by the consumer—are tempting targets for cost reductions. Thus, many purchasing departments are proud of reducing the weight of the board used in cases and trays, without realizing that such reduced-weight materials may cause jams and stops in the equipment that uses them.
In another instance, this one involving a dry product in flexible pouches, the company had specified unique pouch sizes for various products depending on their bulk density in order to reduce their packaging material costs. However, the filling machines they used were difficult to adjust, so they had filling lines for each pouch size, meaning that, on average, many machines were idle. Simply adopting a more common size would have meant better utilization of capital.
Once a line reaches the point that MTBF is low because packaging has inadequate spare capacity, it is difficult to fix. The simplest solution is to reduce processing rate until there is a better balance. Net production may not go down in proportion to the speed reduction because yield should go up. Companies are reluctant to take this step because they usually need production and do not believe that reducing speed might actually help. Almost all other choices, such as installing additional packaging capacity, installing additional surge, or buying better-quality packaging material, are even less palatable, however. Usually, space is at a premium and investing to improve efficiency can be a hard sell.
What to Do About Rework
Practically every food process produces rejected material. How to handle it can be a challenge. Should it be reused? Is rework inevitable? Can it be a hazard? How does it affect yield? There are many different answers.
First, processing material more than once reduces net capacity, but can reduce waste that must be discarded. For instance, a fairly simple example is the trim from cutting shapes from a sheet of dough. This is usually returned to the feed hopper and mixed back into fresh dough. This is a case of “like to like” and also usually happens quickly, so there is not much chance for a change in physical, chemical, or biological properties. It also can usually be done with simple equipment, with little risk of contamination, and for little cost.
--- PAGE BREAK ---
A more challenging example involves material that has been cooked or baked, such as bread, cookies, crackers, breakfast cereals, and pet foods. In these cases, the rejected material may be safe, but is distinctly different from the material from which it was made. It may have reduced functional properties, in which case, it is a relatively inert filler. The fresh ingredients must then do more than usual to suspend the recycled material. For some baked goods, it is believed that some recycled material, with water to hydrate it, can reduce surface stickiness and is seen as a positive benefit.
In other cases, rejected material is used for a different purpose and is not, strictly speaking, rework. For instance, breakfast cereal fines are mixed with salt and seasoning to make a coating product. This seems a good idea until the coating product becomes so successful that there is a shortage of fines, and then good product must be ground to provide enough. The net effect is a reduction in yield of breakfast cereal after all.
It is common practice in some candy and ice cream plants to use rework in darker or strongly flavored products such as licorice in the case of candy or chocolate in the case of ice cream. This poses several challenges. First, potentially perishable rework must be stored properly in sanitary containers and must be properly identified. Second, objective evaluation suggests that mixing flavors hurts quality compared to products made without rework, unless only the same flavors are used together. This may mean holding rework for a while before the opportunity to use it comes around again. Relatively few food materials get better with time, and the rework containers occupy space that might be better used. In these cases, it might be better to discard small quantities of rework than suffer the annoyances of keeping it and managing it properly.
Some ingredients, such as chocolate, are considered so valuable that they cannot be discarded. Off-specification chocolate candy is usually milled through roll stands and mixed back into chocolate that may be used for the same or different products. Because it is impractical to separate the components of a nut, caramel, and chocolate bar, the whole mixture goes back into chocolate. This is why many candy labels warn of peanuts for those concerned about allergens even if peanuts are not otherwise an ingredient in a particular product. It is important to control the amount of recycled material in order to maintain consistency and quality.
Recycling partially or fully processed foods is common in food processing but might deserve reconsideration in light of the space required, record keeping, and quality impact. Effort spent to understand the root cause of off-spec material and prevent its occurrence might provide a higher return than the salvage value realized in some cases. When it is obviously safe and economical to recycle, the rule should be like into like, there should be a rigid upper limit, and the recycled material must be handled safely as a food ingredient.
J. Peter Clark,
Consultant to the Process Industries, Oak Park, Ill.