Silo Holds Mysteries

THIS MONTH'S PUZZLER
We used an old welded stainless steel silo at our Missouri site to store dried brewers grain. During a September shutdown, we installed a new bagger and weigh scale. This prevented us from loading railcars for several days. So, to keep the dryer running during the outage, we cut its output to a rate that would only fill up the silo. The weather was hot but a rainstorm swept through with a cold front. After the storm, we noticed a seam on the silo had split apart, dumping several hundred tons of dried brewers grain. Now, corporate managers, concerned because another plant is planning a similar shutdown and bagger installation soon, want to know why the silo failed. What do you think caused the accident? What should we tell management?

AGE IS AN ISSUE
Evidently, no one bothered to check the outlet temperature of the brewers grain and just slowed the line down without adjusting the airflow and temperature. The grain probably absorbed more energy in the dryer and entered the steel silo hotter than usual.

It was an old silo. The weld was probably stress cracked enough that it would have likely split anyway within the next 6–12 months. The combination of hot grain on the inside (expanding the metal) and cold air/rain on the outside (contracting the metal) probably caused an existing stress fracture to grow until it could no longer handle the load and split.

X-ray the welds on the silos that are going to used during the conversion at the other site. The problem is likely the age of the silo, not the fact that a slight process change sent it over the edge to failure.
John G. Tiessen, applications engineer
Sun Chemical, Northlake, Ill.

FIND THE CAUSE OF STRESS
There are eight forms of corrosion: 1) uniform or general attack; 2) galvanic; 3) crevice; 4) pitting; 5) inter-granular; 6) selective leaching; 7) erosion; and 8) stress. Because of the dry environment in the silo and the lack of erosive forces, the only applicable form of corrosion to consider is stress corrosion.

Stress corrosion is the result of a combination of tensile stress and a specific corrosive medium. A classic example is the failure of brass ammunition cartridges in the tropics after heavy rainfall. There is a parallel in this story of the stainless silo involving high summer temperatures and a sudden temperature change induced by rain. In the case of the silo, any corrosive media must be very slight and the predominant mode of failure is simply mechanical stress.

I would tell corporate management that the failure was most likely due to stress and the cause of the stress will be investigated. I would also speculate the causes of the stress that will be investigated include 1) rapid temperature change in a low strength enclosure and/or 2) stress induced from "flow of solids" rheology phenomenon involving the full hopper and the temperature change.

In regard to "flow of solids" phenomenon, it is common for hopper angles to be improperly designed (not sloped enough) when compared to the angle of internal friction of the solids. The addition of aerator devices to partially fluidize the bed of solids and reduce the angle of internal friction can give far superior results to external vibrators with much less mechanical stress. The sudden temperature change shrinking the metal onto a bed of solids that refused to flow may thus have been the root cause of the mechanical stress.
Mike Gentilcore, chemical process engineering manager
Covidien Imaging Solutions, Hazelwood, Mo.

CHECK THE SILO AND THE PROCESS
We need information to decide if the cause was the silo or the process. For the silo, let's ask several types of questions: 1) How does the failure look? Is there corrosion or embrittlement? Does the weld appear normal? Did failure occur in the weld, at the intersection of weld and welded material, or in the tank wall not impacted by the weld? How do other welds (those not failed) look? 2) Is wall thickness adequate for the density and depth of material stored? 3) Is the silo adequately vented against over-pressure or under-pressure? Does it appear bulged or collapsed? and 4) What is the history of this "old" silo? How many years has it been used? How many fill/empty cycles has it gone through? Has it been overloaded anytime in the past? Is it subject to vibration from attached or other piping or equipment? Has it been left idle for years? When idle, was it kept clean and dry?

As for the process, investigate: 1) Was the grain too wet or too hot? 2) Was there an upset condition due to trying to run the dryer slowly? 3) Did the silo get wet from condensation due to the temperature change or rain intrusion? 4) Was spilled material too heavy for the silo due to being wet or did it ferment and cause gas buildup (then why not vented?)? and 5) Was there a big burp?
Rolfe Hunt, manufacturing technical director
Oil-Dri Corporation of America, Chicago, Ill.

AVOID THE SILO NEXT TIME
This appears to be a materials handling problem. The failure of the old silo could be due to a load factor or a defective weld. Nondestructive testing of the welding, e.g., via x-ray, should be performed. The possible effect of weather on the structure and its contents is probably difficult to ascertain. As far as a similar shutdown, it is suggested that the dried brewer grain be conveyed from the dryer outlet to either bulk 20-ton railcars or truck hopper units to allow production at full capacity; railroad or truck scales can be used to obtain net weights. A diverter valve could be installed in chute work to permit switch over to 50-lb. bags when that system is ready to run.
Robert Drucker, consultant
East Northport, N.Y.

BLAME THE WELDS
It must be defective welding.
Nandan M. Bhandari, engineer
Maple Biotech Pvt. Ltd., Pune, India

WAS THE SILO ALTERED?
Assuming the silo is made of stainless steel, have they cut openings in it? This could cause stress concentration, hydrogen embrittlement and other problems. Inspect the silo to assure that there are proper stiffeners.
B. S. Angadi, engineer
Vijay Engineering Service, Mumbai, India

IT'S EXPANSION
If there was an opening in the silo when the rain storm came through water could have entered the silo and caused the grain to expand.  During June 2008 in Cedar Rapids, IA there was a 500 year flood.  River water got into a soy bean silo and the beans swelled.  The swelling created so much pressure that it caused the concrete silo to fail.
Len Riker, operations superintendent
Penford Products Co., Cedar Rapids, Iowa

HOLD EVERYTHING
There are several facets to this problem. First, is the age of the silo a factor? Second, did running the dryer at a low rate affect the product filling the silo? Did weather affect the silo — was it one storm too many? Was it the mere fact that the silo was run at maximum level? Was something done during the new construction that could have precipitated the collapse?

Clearly all of these questions need to be answered before risking another plant. As with all accident investigations, start with the history. Sometimes information is sketchy. Even with computerized records much is lost or unreliable. Talk to the operators. They'll tell you about the spot repairs, times when the silo creaked or about sudden surges and such; surging could put unusual loads on the thin plates used to construct silos. History may also tell you what products were kept in the silo in the past that might have caused corrosion or erosion damage; it could be that the silo is a retread, a used silo moved to the plant from a junk yard.

The product quality of the dryer at low throughput should be investigated. Introducing moisture could cause the brewers grain to expand. If material was bridged inside, a radial load could put undue loading on old walls.

Weather could be a factor. Perhaps water dripped into an open manway or through an undetected hole in the shell, causing the same effect as mentioned above. It's also possible that a sudden change in atmospheric pressure caused the walls to collapse. Silos are designed for 1–3 psig differential per ASME code requirements for an atmospheric vessel; vacuum is a far greater risk than pressure — 2.5 inches of water column can collapse a tank. Silos tend to be tall with large length-to-diameter ratios. Aerodynamics and weather could have conspired to do in old corroded walls.

Lastly, let's consider new construction. If something was done to the foundation around the tank, or the feeding system at the bottom of the silo was changed, this might have caused a problem. Obviously, if this was done, then management will want to rethink the migration of the new bagger to another silo.
Dirk Willard, senior process engineer
Middough Consultants, Holland, Ohio