ASME - Citrus Waste Pumped Peel Systems

March 12, 1998. Presented by Robert B. Johnston, P.E.

INTRODUCTION
It was five years ago at this conference that Carlos Odio presented a paper on pumped peel systems in Brazil. This was the first time many of us had heard of the concept.

The key advantages claimed were a reduced initial capital investment, along with reduced maintenance and horsepower requirements. These resulted from elimination of many screw conveyors and the peel bin.

I remember many doubts being expressed in discussions following the presentation. The principal concern was in regards to handling the week-end shutdown typical of many Florida processors. Also, the absence of a peel bin meant that juice extraction would have to stop whenever anything went wrong in the feedmill.

Today four of the twenty-five citrus feedmills in Florida use pumped peel. One of these was a greenfield installation that gained the advantage of reduced capital investment. The other three converted from a traditional reaction conveyor system. At least two of these three made the change because they were facing major maintenance and rebuilding costs in aged facilities.

The first two, at Florida Juice and Caulkins Indiantown, were installed by Gumaco using the system described in Odio's presentation. The third system, designed and constructed by Cook Machinery, demonstrated remarkable innovation and advancement in the technology. The latest, at Tropicana in Ft. Pierce, includes further evolution of the pumped peel concept.

Another key development in pumped peel technology came from Cutrale in Araraquara. A paper describing this was presented by Daniel Marques at the 1995 ASME conference.

At that conference I also presented a paper on feedmill technology. In it I mentioned that Dan Vincent's 1940 patent covering peel liming said that three to five minutes reaction time was required. Later in the paper I said that normal practice is to use eighteen minutes. I was aware of this contradiction, so you can imagine my surprise when Marques gave his paper: He said that only four minutes was required and he had a feedmill to prove it.

All four of the pumped peel feedmills in Florida now operate with the short reaction time that Marques described. In fact, Florida Juice and Caulkins Indiantown converted after extended periods of operating difficulties associated with excessive reaction time.

ORIGINAL SYSTEM
To illustrate the technical changes that have occurred, we would like to start with the flow chart in Figure I. This system calls for using recirculating press liquor and molasses as a pumping medium. The presence of this press liquor and molasses, in the ratio of two parts liquid to one part peel, was felt necessary in order to fluidize the peel sufficiently to be pumped.

Note the metal separator tank. This was a relatively large tank that served to mix the peel, lime, press liquor, and molasses. It had a sloped bottom so that tramp metal would not enter the pumps.

The pumps used were progressive cavity pumps. It is interesting to note that today all four Florida plants use progressive cavity pumps built by either Geremia or Netzsch, the two manufacturers referred to in Odio's paper.

The reaction tank was exceptionally large, assuring a reaction time in excess of forty minutes. (The Florida Juice tank was sized for over 60 minutes!) The design called for the peel to be pumped into the bottom of the tank, overflowing from the top. This required operating with a full tank at all times, even during periods when peel flow to the feedmill was reduced.

A major departure from this system was described in Marques' paper. Since only four minutes of reaction time were required, the reaction could be completed in the pipeline running from the metal separator tank to the dewatering screens. This meant that a reaction tank was no longer required. Marques mentioned that the existing Cutrale reaction tank had been converted into a lime silo.

Note that the system requires separation of the pumping medium (press liquor and molasses) from the peel, ahead of the screw presses. The typical Brazilian system used static screens, although some preferred rotating drum screens or shaker tables. A key item is that the pumping medium, already containing d-limonene, flows into the hammermill.

COOK SYSTEM
Let us now take a look at the Cook system (Figure II). This design was installed at the new SunPure feedmill in Avon Park. It was also installed by FMC, under technical license from Cook, at the Parmalat feedmill in Sicily.

One very important change was that only non-recirculating molasses, with absolutely no press liquor, was used as a pumping medium. Also, despite the fact that citrus molasses is more viscous than press liquor, the ratio of molasses to peel was reduced to 1:1.

One fact about the use of molasses instead of press liquor is that improved oil recovery is made possible. This is because press liquor contains all of the d-limonene oil that is recovered in the feedmill. If recirculating liquor is used to pump the peel, there is opportunity for the oil that is present to be absorbed into the albedo of the peel. It is agreed that once oil is absorbed by albedo, it stays there; the action in the screw presses does not remove this oil.

The consequence is that oil absorbed by the albedo goes with the press cake into the peel dryer. Not only is this oil lost, but it is likely to be released, with the exhaust gasses, to the atmosphere. Thus the oil becomes a Volatile Organic Compound (VOC), a focus of the Clean Air Act.

No one disputes this theory. However there are many who do dispute the quantity of oil lost by pumping with recirculating liquor. Comparing oil recovery rates would seem to be the logical way to resolve the question. Unfortunately, conditions at feedmills vary so much that no clear answer has emerged.

There is another subtle, but important reason for pumping with molasses instead of press liquor. Molasses is warm because it comes from the Waste Heat Evaporator (WHE), while press liquor is at ambient temperature, the same as the oranges. As a result, a system pumping with molasses is noticeably warmer than one using press liquor. This difference can amount to 10º C, which does not sound like much.

However, increasing the temperature of a chemical mixture by 10º C cuts in half the time required for a reaction to occur. This is true in the case of the reaction between hydrated lime and citrus peel. Consequently a much faster and more complete reaction occurs, everything else being equal, in a system pumping with warm molasses.

The Cook system had a great many innovations, and, as can be expected with such a system, there were problems to be resolved during start up. One of these led to a change in the design of the mixing tank. It was found that the original tank at SunPure had warm spots and cold spots. These were traced to peel accumulating in the poorly agitated corners of the tank. This peel spoiled and eventually would break loose and get pumped into the screw presses. Screw presses do not operate well with old peel.

A replacement mixing tank was developed by SunPure personnel, and the fabricator, Keller Sales and Engineering, made significant contributions to the final design. (Figure III) Note that the dead spots were eliminated by going to a conical bottom. Internal baffling assured violent agitation, and a clever suction design keeps tramp metal out of the pumps.

There is an interesting detail applicable to any peel processing system. The metal separator used in the Cook plants is the simplest and most effective available. Figure IV shows this device, which is arranged so that molasses is added to the peel by pumping it into the bottom of a small tank through which the peel must pass. The peel enters at the top at one end and exits at the other. Since the molasses is flowing upward, the peel is fluidized as it passes through. The result is that heavy tramp material, including stainless steel, glass, rocks and sand, separates and falls to the bottom of the tank.

Devices similar to this have been installed, with great success, at Orange-co and Cargill Frostproof. Both of these plants have traditional reaction conveyor systems (not pumped peel) where molasses was added either ahead of the shredders or directly into the reaction conveyor. A simple piping change with the addition of a drop-out box has reduced damage to the shredder screens and screw presses at these plants.

Initially the two Cook pumped peel plants (SunPure and Parmalat) used rotary positive displacement pumps. There is a lesson in industrial marketing in what came about. At Parmalat the Italian OMAC pumps have worked fine and are likely to be used in future installations. However at SunPure there were repeated rotor failures. Not only was the pump manufacturer unable to help the situation, but high prices were charged for replacement parts. As a result the pumps were discarded after the first season and replaced with proven Brazilian progressive cavity pumps.

The Cook system uses a different reaction tank design. The peel is pumped into the side of the tank and drawn off at the bottom. This means that an extra set of pumps is required, as compared to the Brazilian system. However it also means that the tank can be operated at any desired level, which translates into any desired reaction time.

Operating at a low level in the reaction tank was found to be advantageous. Florida systems operating with full reaction tanks all had difficulty with screw press operation. After long periods of frustration, the operators at Florida Juice and Caulkins Indiantown traced the difficulty to fermentation that was occurring in the tanks. This seems to occur if there is excessive reaction time, even with good agitation. The result is that today these two plants entirely by-pass their reaction tanks; SunPure operates with a minimal 20% level.

At SunPure, the system was sized for 2,500 boxes per hour (significantly higher throughput can be achieved). At the 2,500 rating there is six minutes reaction time in the pipeline from the hammermill to the feedmill. Additional reaction time is available in the mixing tank and in the reaction tank.

The reaction tank does have an important function at SunPure. Even though the peel goes straight from extraction to the feedmill, without any delay in a peel bin, there are still periods of upsets. When these upsets occur the peel can instantly become un-pressable. At SunPure this condition is overcome by filling the reaction tank until a suitable, pressable, mixture is achieved. Thus the reaction tank is better described as a surge tank.

It is noteworthy that at SunPure there is a provision to add lime solution directly to the reaction/surge tank. This feature is useful when insufficiently limed peel reaches the feedmill.

The point needs to made that pumped peel systems are very sensitive. You will recall that our original worries were about weekly start-ups and having to shut down extraction if problems occurred in the feedmill. It is fair to say that it is the sensitivity to change that has caused far more problems to operators than the problems we anticipated.

By sensitive we mean that the presses stop working and operations become near impossible. There are many normal upsets that cause immediate problems in the pumped peel systems: a slug of CIP water, a load of cull fruit, trash water in the molasses, peel over a few hours old on a hot day, a jump in molasses pH, a skip in the limer. All of these lead to wet peel going into the dryer, which has resulted in multiple cases of smoldering peel in the dryer.

Returning to the SunPure system, another innovation is evident in how the peel and pumping medium are separated ahead of the screw presses. Instead of open screening devices, Cook chose to use closed pre-presses. These machines have limited compressive characteristics, so they perform a "soft" press on the peel. These have proven noticeably more effective than screens in separating free liquid from the peel. As a result, the screw presses that follow give better performance.

One feature of the SunPure system is the use of spent caustic as a replacement for lime. In the 1950's, working at Citrus World, R. W. Kilburn showed that spent caustic could be substituted, up to 50%, for lime. In oversimplified chemistry, in order to prepare peel so that it can be pressed it is necessary to break down the pectins and to raise the pH: the spent caustic will raise the pH, while it takes lime to break down the pectin. Cook chose to take advantage of this, using 10% to 20% caustic in the lime slurry. Lime and spent caustic are mixed, in batches, and then pumped to the hammermill.

One advantage of this is that, since less lime is used, there is less lime going to the Waste Heat Evaporator with the press liquor. As a result the formation of calcium citrate is reduced and the WHE works more efficiently for longer periods.

One aspect of this technology is difficulty of control. In a reaction conveyor system the addition of lime to the peel is set by matching the speed of the limer screw to the speed of the peel bin take-out screw. The operators adjust this system by measuring pH. In a pumped peel system it is more difficult to match the lime flow to the peel flow since the peel flow is not evident to the feedmill operators. Thus the operators are all the more dependent on pH as a means of controlling lime addition. Unfortunately, pH meters tend to be unreliable, leaving the operator uncertain about how much lime is required at any given moment.

One concern frequently voiced about pumped peel systems is that, since there is no peel bin, extraction will have to be shut down if something goes wrong in the feedmill. In fact, some of the Brazilian and most of the Florida pumped peel plants do have peel bins. The more practical solution is seen at SunPure where there is no peel bin as such. Instead as a first alternative they have provision to divert peel to a dump pad. From there the peel (both their own and peel from other processors that they run on occasion) can be put back into the feedmill flow with a front end loader. Their second alternative is to divert the peel to a small loading silo from which it can be trucked off-site. This system has worked well.

There are other innovations at the SunPure feedmill unrelated to pumped peel. The principal one is found in the WHE which is a five stage, three effect evaporator with advanced technology in the oil stripping area. This 80 WHE is matched to a 40 dryer which has resulted in excellent thermal efficiency.

TROPICANA SYSTEM
Figure V shows the system that went into operation at Tropicana Ft. Pierce last year. It can be described as a progression in pumped peel technology. It is characterized by gross simplification: the mixing tank is designed to provide surge capacity; the reaction tank is eliminated; and the pre-presses or dewatering screens are also eliminated.

A key to this simple system is pumping with a low ratio of liquid to peel. At Tropicana only one part molasses is added to three parts peel. To our surprise, the progressive cavity pump handles this thick mixture with no evident difficulty.

Flow through the pipeline is mostly laminar; there is little turbulence. Therefore the mixing of the lime with the peel must be thorough and complete before the peel enters the pump. At Tropicana the dry lime is added at the peel bin take-out screw. Molasses is added shortly afterwards, ahead of the hammermills. At times the mixing tank is operated at such a low level that very little mixing occurs in the tank. We were surprised that this relatively limited and simple mixing of peel and lime has proven quite adequate for achieving a proper reaction.

Operating with the peel from 3,000 boxes per hour, there is about one minute reaction time between the limer and the hammer mill. Six minutes are available if the mixing tank is full. Finally, there is one minute reaction time in the pipeline.

The mixing tank design is worthy of note. The Keller units at SunPure and Tropicana feature vertical internal baffles that work in conjunction with three tiers of center-mounted rotating paddles. Dead spots are minimal. The suction pipe requires the fluid to make a U-turn, assuring that tramp metal is effectively separated in the bottom, ahead of the pumps.

The original Tropicana system included the Vincent pre- presses shown in Figure VI. These were hard-piped directly from the peel pump. This means that the peel, from the time it enters the pump, is totally enclosed, completely preventing the emission of any VOC's.

As the ratio of molasses to peel was reduced, it was found that the pre-presses were not necessary. The main presses alone were found capable of removing the pumped fluid as well as doing their normal job of pressing the peel. The pre-presses were removed; the system as it is today is shown in Figure VII. Note that the peel is hard-piped to the presses.

It was found necessary to protect the hard-piped system from over-pressure. This has been achieved by the addition of a pressure relief valve, shown in Figure VIII.

CONTROL
Under steady full-load conditions, there are few problems controlling either traditional or pumped peel systems. However with pumped peel systems operators need to be more alert and learn new tricks to handle upsets and start-ups.

The upsets have been previously enumerated: a slug of CIP water, a load of cull fruit, trash water in the molasses, peel over a few hours old on a hot day, a drop in molasses pH, a skip in the limer.

In a traditional system these present a minor inconvenience for several reasons. There is usually the opportunity to mix some fresh peel from the peel bin with the bad peel. Peel flow is easily controllable by changing the speed of the peel bin take-out screw. Reacted peel color is readily observed. Lime flow is readily adjusted, sometimes by dumping bags directly into the reaction conveyor.

These are not characteristics of a pumped peel system. As a rule, the operator can observe only pH, Brix and the discharge cake from his peel presses. Thus the operator must learn to spot what is causing the upset and know what to do about it. His choices are largely limited to adjusting the lime flow and slowing his pumps so as to fill a surge tank. Filling the surge tank allows both mixing of new material with the material causing the problem as well as slowing the input to the presses.

One change to pumped peel systems that has been proposed is the addition of a weigh belt. A weigh belt would provide the operator with an indication of the tonnage of peel entering the system. This would be helpful in situations where extractor lines are started or stopped. It would give the operator knowledge of where his limer control should be set.

Originally there were serious concerns about start-up operations. Plants that shut down each weekend face a Monday start-up with water, sour press liquor, or cold molasses. However, many plants, both in Brazil and Florida, routinely get through their "Monday bad hair day." It takes careful attention by the operators, and efficiency is less than optimal, but the difficulties have proven readily surmountable.

EVALUATION
Despite the difficulties of start-up and control, there are strong reasons favoring the pumped peel system. The simplicity of the Tropicana system has obvious advantages of reduced maintenance, space requirement, and capital investment.

This same simplicity has important implications from the standpoint of VOC control. In a pumped peel system the peel is fully enclosed and captured from the inlet to the hammer mill all of the way through the screw presses. This eliminates a host of potential VOC emission points; the equation is simplified.

Older traditional feedmills experience a great deal of maintenance. This is reflected both in operating costs and, more importantly, downtime. Screw conveyors, and especially the reaction conveyor, are notorious for failing and interrupting feedmill operations. It is not uncommon for these situations to cause a halt in juice extraction operations. Because of this condition it becomes financially attractive to convert an older system to the pumped peel technology.

Nevertheless we should note that the presses, dryer, WHE, pellet mill, and pellet loading remain unchanged in a pumped peel feedmill. Consequently it cannot be said that fewer maintenance personnel and operators are characteristic of pumped peel feedmills.

SUMMARY
The technical progress in Florida, in pumped peel systems, has come as much from the citrus processing community as it has from the equipment suppliers. The early adaptors were the CEO's of their organizations: Ron Grigsby at Florida Juice and Roger Beret at Caulkins Indiantown. It is notable that both men were relative newcomers at the time the key decisions were made. It was their decisions that encouraged other processors to install pumped peel systems.

At SunPure it was Hadi Lashkajani who made the choice to go with the many innovations offered by Cook Machinery. Vice President Donald Dawson and Feedmill Manager Mac Greene stand out for having spent the innumerable hours required to make the system the success it is today.

Dave Van Etten, Plant Manager, had the vision that resulted in the Tropicana system. He heard Marques' presentation, went to Araraquara to see it, and came back convinced it offered many advantages in his plant. He is quick to give credit to Chris Sutherland, Instrumentation Manager, who had the understanding of what it would take to make it work on a day-to-day basis.

To summarize the technical progression that has occurred over the last five years: The Cutrale work with short reaction time allowed improvements at Florida Juice and Caulkins Indiantown, while it led into the innovation at SunPure and Tropicana. Ralph Cook first introduced concepts that will be fundamental for years to come: the avoidance of recirculating pumping fluid; bottom discharge reaction tanks; spent caustic as a lime substitute; using pre-presses instead of dewatering screens. At Tropicana the key new features are pumping with a limited amount of molasses, eliminating the reaction tank, and hard-piping to the presses. It remains to be seen what the next system will be like.

THE NEXT FIVE YEARS
A forecast of the future should take into consideration other technical innovations in feedmilling technology. Items of note include:

Work directed by Benedito Jorge at Citrosuco has led to the development of a single pre-press that will handle the pumped peel coming from fifty FMC extractors.

Vincent pre-presses have been added to the pumped peel systems at both Florida Juice and Caulkins Indiantown. These have improved the performance of the main presses at both plants.

At Cutrale's Leesburg feedmill, the use of twin screw superchargers, with overlapping flights, has led to improvements in press performance. This has important implications in double pressing operations. See Figure IX.

The Fiber Filter, shown in Figure X, is being developed to clean-up of press liquor. This will allow WHE's and pumps to operate with higher Brix molasses than was previously possible in many plants. This results in reduced WHE fouling and improved thermal efficiency.

The use of twin circuits in the WHE, to produce higher Brix molasses for second pressing, is likely to result in significant improvements in overall thermal efficiency. One Brazilian plant is using this principle with solid results.

Some Brazilian processors have found that eucalyptus stands are more effective for disposing of wastewater than orange groves. This can impact the demands made on the WHE.

The bagged peel system, described at the most recent Citrus Short Course, eliminates the feedmill dryer altogether. Air pollution regulations in California have led to the adoption of this bagging process. Either VOC regulations or the low price of pelleted citrus peel could bring it about in Florida.

Dreyfus in Winter Garden has now begun feedmill operations using a unique Vincent screw press. Their new press features a screw design originally developed for alcohol extraction in soybean plants. In addition, the press makes use of profile bar screens, as opposed to the standard 3/32" perforated screen material. Its performance on limed orange peel is being tested.

This is a remarkable list of technical innovations and experiments. It gives us assurance that the next five years will be as eventful as the last five.