Gerald Zamiski, Ph.D., P.E., of Vollmer-Gray Engineering Laboratories conducted the United States Consumer Products Safety Commissions study on particular Nox rod equipped, horizontal forced air furnaces manufactured by Consolidated Industries. Here, Dr. Zamiski demonstrates a burner assembly for a film crew reporting on the fire and life safety hazards posed by these furnaces.

This article is from the March 2001 edition of "The California Fire-Arson Investigator."

Vollmer-Gray Engineering Laboratories, Inc., in Long Beach, California was hired by the Consumer Product Safety Commission (CPSC) in Washington, DC in 1997 to investigate the failure modes for NOx rod equipped, horizontal forced air furnaces manufactured by Consolidated Industries. The project involved dissection of 96 furnaces and thousands of hours of inspection, analysis and testing.

This engineer has examined close to a thousand failed and non-failed Consolidated furnaces which includes close to 50 from residential fires. The Laboratories has been hired by dozens of builders and developers as consultants. Field testing of hundreds of homes has produced statistical data on the operational parameters in which the furnaces operate under. The combination of the CPSC research, furnace examinations, and field studies produced findings as to the cause of the furnace failures.

INTRODUCTION

The analysis tasks were designed to yield information regarding failure characteristics of the 1983 through 1992 HAC and HCC horizontal forced air Consolidated furnaces with California emissions "NOx rods". The furnaces utilized for the CPSC report consisted of HAC/HCC models of size 60,000 BTUH (3 burner), 75,000 BTUH (3 burner), and 100,000 BTUH (4 burner). Prior to removal, field operational parameters were observed and measured. This data included: model number, serial number, BTUH rating, limit switch settings, temperature rise, external static pressure, clock time for gas flow, and gas valve pressure.

The furnaces were visually examined and disassembled. The disassembly involved removal of the burner and all lower panels so as to fully expose the bottom of the heat exchanger. Each burner assembly was microscopically examined. The burner folds were numbered 1 through 31 beginning at the manifold end. NOx rod clips were located at folds 1, 8, 16, 24, and 31. As a reference, fold number 24 (with clip) corresponded approximately to an exchanger position 18" from the return end.

The burners and exchangers were categorized separately based on visual characteristics representing degree of use, state of degradation and progression of failure. Metallurgical samples were then removed from the selected units. Metallurgical mounts were made from 85 samples: 38 heat exchanger and 47 burner samples (includes 8 "full" mounts defined as the NOx rod, clip and burner fold in relative position as found during usage). Scanning Electron Microscopy (SEM) samples totaled 32, which included 8 exchangers and 24 burner folds/ ports. Energy Dispersive X-ray Spectroscopy (EDS) analysis was used on 8 burners and 8 exchangers. Bulk chemistry was performed on 30 samples: 10 burners, 9 exchangers, 7 rods, and 4 clips. Micro-hardness measurements were made on 48 samples: 19 exchangers, 21 burners, 4 rods, and 4 clips. The above data, readings, observations and characteristics were analyzed and summarized.

HISTORICAL INFORMATION

The Horizontal Forced Air Furnace model "HAC" design began in 1979 and was first sold in 1983. The material used for the burner tubes is T140 aluminized steel. The burner ports are made by a "progressive die method" where dies punch holes in the steel. The burner tube is formed by a single piece of steel which is folded and crimped to form the ports.

The first NOx rod-equipped furnace for California was sold in 1984/5. The rods were simply added to the existing burner design with "no changes to the furnace ... other than clips." Prototype testing/evaluation notes of the NOx rod-equipped units tested between 1982 and 1984 revealed that the rods exhibited a hot spot or region of higher temperature. In this hot spot region, the "rods glowed red." The non-NOx rod HAC design was fully tested by the American Gas Association (AGA) in the summer of 1983. The NOx rod unit was not fully tested, but rather given "abbreviated" testing due to the "relatively small change." The heat exchangers were originally made out of cold rolled steel.

The typical model numbers for Consolidated furnaces, HCCIOONDSRX, display the following information:

HCC stands for Horizontal, Model C and Cold rolled steel exchanger. (Models with aluminized steel heat exchangers used the designation "A" instead of "C".)
The number 100 is the KBTUH rating.
N is natural gas
DS is control type
R is for relay for blower
X is for NOx rods.
SUMMARY

Chemical analysis identified the materials utilized for the various components. The burner base material, underneath the aluminized coating, and the cold rolled steel (CRS) exchanger were both made of 1005-1006 low carbon steel. The NOx rods and clips were made of 330 stainless steel.

SEM analysis provided the following observations of the results of the die stamping and fold forming operations. The folds displayed a number of significant manufacturing flaws which developed from the die stamping and fold forming operations. The comers of the stamped port opening were sharp. A shear burr was formed to one side of the stamped edges. Upon forming the burner tube, and as the folds were made, the sharp comers and shear burrs were found to be prone to cracking. The top of the fold showed cracks in the aluminum coating. Polished and etched metallographic sections showed that the fold bend radii were below the minimum allowable value by a substantial amount. The minimum radius should have been 0.015". The actual values averaged 0.0055", with some radii as low as .0030". The result of the extreme fold bend radii was cracking and spalling of the aluminized coating along the outer radius of the fold, and heavy microstructural damage and cracking at the inner radius of the fold. The lower inner fold base corners, adjacent to the port, had very sharp comer radii (as low as 0.004") and exhibited cracking on most of the folds examined.

The burner degradation occurred at an advanced rate relative to the exchanger breakdown. The first sign/step of the failure process for the burner is fold oxidation and cracking. SEM analysis showed that the initial cracking due to service induced degradation occurred across i) the top of the fold, propagating from the splits in the shear burr and ii) across the lower inner fold propagating from the lower base comer cracks, which were induced by the crimping operation during manufacturing. The burner fold failure process involves preferential oxide development at sites with manufacturing flaws. As the breakdown of the fold progresses, the top of the fold begins to lose substantial portions of material due to oxidation, cracking, and spalling of the oxide from thermal cycling and further oxidation, as shown in Figure 3. By Category 5, folds in the heaviest damaged burner zones star to become fully fractured due to the oxidation-induced process noted above. By Category 5+, oxidation induced damage spreads to the top of the port walls. Th level of temperature and the thermal cycling causes oxidation-induced consumption of the folds and the distortion to the port geometry. Category 6 burners show full fold fracture and consumption and substantial port opening in the hottest zones.

The majority (85.5 %) of the 100,000 BTU burners with NOx rods had at least some degree of burner fold cracking. One quarter (25.9%) of the burners were in the final stage of full failure. For the non-NOx rod units 100% of the burners had no cracking. This is in clear statistical contrast to the NOx to equipped units. Fold 24 along burner row 4 and 5 were the first folds to crack, and the folds with the highest occurrence of cracking. The zone of maximum fold degradation ranged from folds 20 through 28 along rows 3 through 6. Over 50% of the burner assemblies had a visible "hot spot" discoloration zone on the panel below the burner. This hot spot averaged 6.6" wide, and ranged from folds 13 to 31. Over 50% of the NOx rods displayed warpage and 44% displayed residues from combustion. Several rods were found to have sustained massive consumption. Chemical analysis showed that the consumed rods were made from 302 stainless steel, a lower grade of stainless steel with 27% less nickel than 330 stainless steel. Fold 24 of the burner corresponded to a point 18" along the 24" exchanger. The exchangers displayed a local hot spot where metallurgical breakdown occurs. The hot spot was centered at 18" along the exchanger, 1" above the bottom weld, and at the four central modules. The hot spots were noted in 14.7% of the exchangers. Five of the exchangers exhibited more than one hot spot with one even having three such zones. The heat exchanger hot spot breakdown is the result of a similar oxidation-induced failure, as shown in Figure 4. The level of temperature along with the cycling of the furnace causes preferential oxidation, cracking, and spalling of the oxide, and further oxide growth. The typical exchanger had cracking on the supply and/or return expansion joints. The average transverse crack on the expansion joints were 7.1 " long on the supply side and 9.2" on the return side. The average longitudinal crack length was 0.9" for the Utilizing the method of least squares and statistical linear correlation showed that there was no correlation between the degree of burner or exchanger breakdown and any of the field operational parameters. The data in most cases resembled "scatter".

CONCLUSIONS

The manufacturing process utilized by Consolidated for fabrication of the burner causes several flaws which are major contributors to the burner and heat exchanger failure process. The die stamped port openings have corners with insufficient corner radii. Upon folding and crimping, the inner comers crack. The fold forming radius is more acute than the minimum recommended radius. As a result, the aluminum coating on the new, as-manufactured fold cracks, spalls, and loses adherence. The shear burr on each end of the fold splits and cracks during the folding process. Due to the manufacturing damage to the aluminum coating at the folds, the thermal rating of the burner metal decreases from a maximum of 1250 degrees F (ANSI Z21.47a allows 1100 degrees F) for the aluminized material to a maximum of 1000 degrees F (Z21.47a allows 900 degrees F) due to the now "exposed" low carbon steel base material.

During furnace operation, the NOx rods raise the temperatures to which the burners are exposed. With exposure to the NOx rod elevated temperatures, an oxidation-induced breakdown of the burner folds occurs. Preferential oxidation and crack propagation occurs at the shear burr cracks and the fold inner comer cracks. The folds eventually are consumed from excessive temperature and a thermal cycling-induced oxidation process.

The NOx rod implementation on this burner causes temperatures to exceed that allowable for the materials. It must be noted that, to this date, no non-NOx rod burner has been examined which shows any burner fold cracking and/or degradation. Thus, fold flaws from manufacturing are present on all burners, but the operating temperatures on units without NOx rods are insufficient to cause burner breakdown.

Once a sufficient number of burner folds have fully fractured/consumed, the thermal environment causes the burner ports to distort in an open mode. Eventually, a number of ports open and combine into one large "port" which can include as many as 6 or more of the original ports. The resulting condition allows a large, "lazy" flame to exist. During the fold breakdown and into the port opening stages, the heat exchanger sustains a localized overheating which generates cracking and eventual rupture. The final stage in the heat exchanger failure process allows blower air to exit the exchanger rupture, impinging on the large lazy flame and forcing it downward and laterally. The flame can be forced out the side burner opening and ignite adjacent combustibles. During the same sequence, the flame and heat at the burners are blown downward into the panels below the burner. This increases the temperatures at the furnace to wood flooring interface, and can lead to a pyrolytic decomposition.

The information in this section is based on deposition testimony and documents produced in private litigation, documents provided by Consolidated to CPSC, and personal experience and observations.