FCCU Operation Monitoring and Problem Diagnosis -
by Bob Flanders
edited by Ronald G. McClung
This Catalyst Report is the first of two dealing with the fundamentals of monitoring an FCC unit operation. The wisdom in these two articles comes from the late Bob Flanders, a learned gentleman who spent many years in the petroleum refining industry, and retired from the Chevron Corporation several years ago. Some of Bob's comments are specific to Model IV unit design and have been so noted in the text. Many readers may recognize Bob's writing style as the same as his speaking style. Bob wanted this published so that others might benefit from his many years of experience. To that end, the two Reports are divided into monitoring performance where catalyst may have an impact and then general considerations related to equipment, feed and product analysis.
Many times the operator reports "The cracker has no punch". Also nothing seems to explain it. Conversion loss can be related to a problem with the feed, catalyst, or process conditions. More often than not conversion loss is a mechanical thing caused by the operator himself.
First, suspect that some of the feed is not getting to the reactor. Check the bottoms-feed exchanger for leaks. In most plants the feed side is the high pressure side. A simple qualitative test is to introduce some nitrogen from a cylinder into the feed downstream of the pump. If a sample of fractionator bottoms bubbles on standing, the bubble is probably nitrogen. You have a leaker.
Second, if the process design provides an emergency by-pass of reactor feed to the fractionator, try the "hand pyrometer" test. One should be able to place a bare hand on the by-pass line down stream of a closed valve. If the bypass line is too hot for touching, you have a leaking bypass value resulting in low conversion.
Loss of catalyst activity could mean loss in conversion. A frequent cause is sodium in the feed either from caustic or brine. Incidentally, sodium in the feed will also reduce the activity of CO oxidation promoter catalyst. Residuum in the feed will often contain sodium particularly where caustic is used to control corrosion in the crude units.
Vanadium in the feed has been a problem with respect to catalyst deactivation particularly where sodium is also present. Serious catalyst deactivation occurs from vanadium contamination where regenerator temperature exceeds 1300 °F and all the CO is combusted.
It is well established that nitrogen bases are temporary poisons for the cracking reaction. Based on laboratory tests, aromatic compounds are worse than paraffins as nitrogen sources. Aliphatic nitrogen compounds crack readily to ammonia in FCC reactors.
The feed hydrofiner is of little use at low hydrogen uptake because nonbasic nitrogen compounds can be made basic. At 200 to 300 ft3/bbl hydrogen uptake hydrofined feed is more refractory than raw feed to an FCC reactor.
Carbon on regenerated catalyst (CRC) is another temporary conversion suppressor. The effects on conversion loss of CRC are non-linear, decreasing in effect as the CRC rises. For example, over the range of 0.1 to 0.3 wt. % CRC, commercial unit conversion loss will be approximately 1% per 0.1 % CRC.
If CRC is a problem, try to raise excess O2 or reduce the concentration of feed coke precursors by, for example, crude still cutpoint adjustment or dilution of the feed with LCO, at constant total feed).
Catalyst choice and makeup rate contribute to equilibrium catalyst activity, which in part determines conversion of feed. Some of todays catalyst are too active for easy-to-crack feeds. In most cases three MAT conversion units will shift actual unit conversion about 1% to feed. The effect is non-linear. The higher the conversion the less MAT changes affect conversion.
Catalyst choice and makeup rate are listed last in consideration of conversion loss because timewise, catalyst policy is a long range item. No quick solution will come from a change in catalyst policy. Coke and gas are conversion products but are seldom money makers. Conversion per se is rarely a useful operating parameter.
Hydrogen Make Control
Competing dehydrogenation reactions account for the net hydrogen made in the FCC unit. Dehydrogenation is time dependent and independent of cracking activity. The worst offenders are nickel and vanadium. These metals at zero valence have significant dehydrogenation activity. This suggests a temporary solution to a surging compressor, reduce reactor or riser residence time by steam dilution. If there isn't enough time in the reactor/riser to reduce the metals the hydrogen problem will be reduced. Tramp iron circulating with the catalyst will increase the sulfur oxide level in the flue gas and reduce, proportionately, the H2S in the gas.
Acceptable hydrogen make is that rate which does not interfere with wet gas compressor operation or restrict FCC capacity. A plant making 20-40 ft3/bbl of feed is doing well. Twice that rate is bad but three times that rate will probably mean a surging compressor which needs immediate attention. As solutions, it was already suggested to introduce steam dilution to the reactor riser. It may help to raise excess oxygen in the regenerator to further oxidize contaminant metals. The last solution is discard and rapid replacement of circulating catalyst. Use of antimony or bismuth additives has known capacity to tie up contaminant nickel in a form which has low dehydrogenation activity. Neither purging of catalyst or use of passivators are fast enough to help stop a compressor from surging. This requires quick emergency steps.
Rough Catalyst Circulation
Phenomena sometimes called 'bumps", "slides", "blips", etc. can indicate bridging and collapsing of catalyst flow in lines and vessels. These are identified by sudden shifts in bed level (Model IV specific) either in the reactor or the regenerator-- sometimes the stripper.
Suspect either a plugged aeration tap, or following a shutdown, a missing RO (restriction orifice). Process designers specify aeration rates. Use them unless there has been a change in catalyst density or circulation rate. If catalyst density or circulation rate changes, aeration rate may need to be adjusted. Monitor each aeration tap with a machinist's stethoscope. This instrument was first used by FCC operators to make sure steam traps were working properly. They can be useful in finding plugged or missing RO's. More often than not operators tend to over aerate. With a machinists stethoscope a missing RO sounds like a waterfall. A plugged one is silent. A properly sized one has a high pitched scream typical of sonic velocity. Poor service should be expected from any RO sized smaller than 5/64 inches. These sizes are easily plugged with line scale.
Another suspect in rough circulation is the size distribution of the circulating catalyst. Some plants are more sensitive to catalyst size balance than others. All fines or all coarse sized particles are very difficult to fluidize. A size range of 6 -12% 40 microns and finer together with lower 80 microns and larger is preferred. In a fines deficient system, recycle of the material caught in the first stage of an electrostatic precipitator will give a sudden burst to catalyst circulation. The effect is short lived without continued recycle of fines. It is questionable that purchase of fresh catalyst spiked with extra fines is a winning proposition. The plant has already shown that it cannot contain fines whether purchased or internally generated.
One scheme to permit small changes in aeration rate without a changeout of RO's is to install a pressure regulator on the steam line to the aeration manifold. U-bends flowing either vertically downward or upward (usually specific to model IV's) should have separate controllers on each transfer line.
Make sure aeration lines containing steam are well insulated. If any free water flashes in a metal line, fluidization will be upset, but worse, this can lead to line metal failure because of stress corrosion or metal fatigue.
Check the delta P between taps in catalyst lines. A zone of zero pressure drop can either be void or a dense pack. In either event circulation cannot be good without design pressure balance. Radiography is useful in estimating catalyst density in transfer lines, but for some this requires an outside contractor. Blast steam is a last resort.
Some operators are inclined to use blast steam as a cure all. Any sudden burst of steam should only be used after the steam is bled to dryness. Blast steam is a catalyst attriter, and wet steam can give a dangerous pressure surge which could lead to a flow reversal, or worse, a line failure.
Gasoline Octane Loss
The most frequent cause of loss in FCC gasoline octane number is a feed quality change. Paraffins or feed hydrofining of some or all of fresh feed will reduce octane number of FCC naphtha relative to that generated from naphthenic stocks. Any straight run heavy gasoline either from crude or generated in feed hydrofining will go through the reactor at about 10% conversion and be about 35 octane number. (Editorial Note: This naphtha conversion is typical for a (bed) cracker. For short contact time riser cracking, this conversion will be substantially reduced.) If FCC gasoline octane number is improved by undercutting endpoint, suspect gasoline boiling range material in the feed. An ASTM D-86 distillation of fresh feed is recommended to alert the operator to the gasoline overlap problem. Upgrading of low octane gasoline stocks in an FCC reactor will not be very attractive. The inclusion of thermal stock from a coker or a visbreaker will lead to reduced conversion and gasoline octane unless the stock is deeply hydrofined. The reactor cracking temperature strongly influences the olefin content of catalytic naphtha and conversion level very much influences the aromatic content. The content of both olefins and aromatic directly effect octane.
An octane distribution analysis can be very useful in evaluating changes in feed type or reactor conditions or catalyst choice. Figures 1 and 2 show inspections of laboratory cuts of gasolines from two plants operating on different feeds. The "A" plant was cracking a naphthenic gas oil once through at high conversion. The feed to the other plant, "B" was from a paraffinic crude operating in recycle mode with some residuum in the feed blend at low conversion. Note the very low heavy gasoline octanes reported for Plant B. The low conversion rate of feed results from sodium contamination of the catalyst resulting from use of caustic for corrosion control of the crude still.
Heavy gasoline octanes tend to be low with paraffinic feeds because the FCC does not do a very good job in dehyrocyclizing high boiling paraffins. Even a catalytic reformer struggles with high boiling paraffins but it does much better than an FCC. An attempt was made to explain poor heavy gasoline octanes for unit 'B' based on inadvertent insertion of extraneous gasoline fragments in Plant B feed. The FCC operators insisted that gasoline fragments were not present to cause this low backend octane.
High catalyst loss not only costs up front, but can lead to excessive erosion of equipment. Operators as a matter of custom blame the catalyst vendor for any unexpectedly high catalyst losses. Many sudden increases and many long term gradual increases in catalyst loss are finally identified with process or mechanical problems.
First, identify whether the catalyst is leaving via the reactor or the regenerator. Look at the size distribution of particles in the flue stack, the fractionator, and equilibrium catalyst. The flue stack can be tested with the Anderson Sampler. The catalyst in the fractionator bottoms can be filtered for a screen analysis. It takes a large sample (2 gallons) but your catalyst supplier can handle the test and the filtration.
Large particle sizes in the flue gas or fractionator suggest plenum or cyclone failure. Mostly fines may mean a plugged or partially plugged secondary trickle valve which for many units can easily account for 10 or more tons/day of incremental catalyst loss.
High fines in the flue stack together with low fines in the circulating catalyst could mean mechanical distress near the grid or air distributor. If the pressure drop is low in the air inlet system, the resulting air rate will lead to an excessive superficial velocity in the regenerator. This high air rate will then cause fines to be preferentially lost from the dense phase.
Normally in this situation, CRC tends to rise because of poor air-catalyst mixing. The catalyst may have a salt and pepper look.
High fines content of the fractionator bottoms stream can mean reactor cyclone distress or bypassing in the stripper. Look for hot spots or cold spots on the stripper shell using a hand pyrometer or infrared photography. Another source of high fines is the sudden development of a catalyst attriter. Examples are a broken steam distributor in the stripper or undetected use of a blast steam nozzle.
Another possible cause of high catalyst loss, particularly from a reactor cyclone system, is the choice of the type of secondary trickle valve. Figure 3 and 4 illustrate two different trickle valve designs, one with a vertically oriented flapper and one with a horizontal flapper. The design of Figure 3, with the vertical flapper is recommended for use submerged in a dense phase. Plants with riser reactors will have poor success with this design of valve. The horizontal flapper design, Figure 4 has a counterweight and is recommended for use in a dilute phase.
Last, ask the catalyst vendor to recheck the quality of recent shipments. Most suppliers retain samples of fresh catalyst shipments for some period of time.