This article focuses on the collapse of an excavating machine’s seat pedestal at the TU Electric Oak Hill strip mine near Tatum, Texas. It became serious when the excavator operator reported that the accident had injured his lower back. The strip mine’s insurance company brought a faulty product lawsuit against the manufacturer of the excavator. The seat assembly was modeled in an upright position using three-dimensional beam, plate/shell, and solid brick elements. The beam and plate/shell elements were used to represent the steel base of the seat and were defined using the material properties of steel from the software’s Material Library Manager. Giesen followed the event step-by-step on his computer screen and watched the force operating at points on the body in real time. It was remarkable to see the reverberation in very high mechanical frequency rippling up and down the backbone. The report of Giesen’s destructive examination and Mechanical Event Simulation results was one of many considerations in the subsequent settlement of the lawsuit.
It started out as an ordinary day at the TU Electric Oak Hill strip mine near Tatum, Texas. It became less than ordinary when the excavating machine’s seat pedestal collapsed. It became serious when the excavator operator reported that the accident had injured his lower back.
The strip mine’s insurance company brought a faulty-product lawsuit against the manufacturer of the excavator. So it eventually became the task of the Henderson, Texas, law firm of Wellborn, Houston, Adkinson, Mann, Sadler & Hill LLP to prove that the mechanical failure of the seat caused the operator’s lower back injury.
Raising a New Question
When it took over the case, the firm contacted Herman M. Giesen, a Dallas professional engineer and engineering consultant who had been researching the cause of the failure. Giesen had been working first for the insurance company itself, and then for a lawyer who had taken the case before it came to Wellborn, Houston.
This article was prepared by staff writers in collaboration with outside contributors.
By that time, Giesen, an ASME member, had reviewed design drawings and conducted a destructive test of an excavator seat pedestal of the same part number. Mark Mann, the lawyer who handled the case, raised a new question that sent Giesen seeking a solution: What was the impact suffered by the operator’s lower spinal column when his seat collapsed?
A Varied Background
Giesen was associated with the U.S. Air Force for 15 years, then became involved in the private sector aerospace industry. His consulting business has given him extensive experience in forensic engineering, along with a long history of systems engineering and mechanical systems design.
Pursuing an answer to Mann’s question led Giesen to his first practical experience with finite element analysis. He eventually would use Mechanical Event Simulation software from Algor Inc. to calculate and present a visual reconstruction of impact forces, stresses, and oscillating movement.
Giesen’s research into the cause of the failure focused on the seat’s pedestal assembly, based on a witness’s report that the seat had apparently failed in the area of the torsion assembly axle. The seat of the excavator rests on a pedestal assembly that includes an adjustable torsion spring scissors mechanism. This mechanism lets the operator adjust the seat to a proper height to reach two pedals and a hand manipulator comfortably.
Giesen was unable to study the seat pedestal actually involved in the accident, although he tried to recover the seat at the mine site. He thought that it was buried there, but it had been discarded, and a search of the site failed to recover it.
Thus, instead of the original, Giesen decided to acquire two seat pedestals of the same model and the seat pedestal design drawings through the legal discovery process.
By reviewing the design drawings, Giesen detected two important facts about the design: The bearings that support the axles overhang the axle roots, and there are no fillets at the roots of the axles. The effect of the first fact is to create a significantly overhung load, which stresses the root heavily under a bending load. This stress concentration at the root is further amplified by the second fact—the absence of root fillets.
“Failure to provide axle root fillets was a design flaw and was the root cause of the failure and the operator’s injury,” Giesen wrote in his final report. “Had appropriate fillets been provided, the event most probably would not have occurred. Given these design flaws, chronic cyclical and vibrational stress are likely to cause fatigue cracks to develop, propagate, and cause failure.”
As Giesen moved forward with the destructive disassembly and examination of the sample seat pedestal to confirm that there were no fillets at the axle roots as manufactured, he also discovered that a hole called for in the plans to support the axle was tapered and ragged, possibly because the hole was punched in manufacturing.
Design and manufacturing flaws were cited as causal factors in the failure of the seat pedestal.
“The tapered axle hole would have allowed for an axle root fillet radius of approximately 0.02 inch,” Giesen wrote. “A fillet of that size would have significantly reduced the stress concentration factor and, hence, the likelihood of the failure. Alternatively, a non-tapered hole would have better supported the axle root as machined. Either way, it was unambiguously clear that no relief had been specified in the machining of the axle root.”
By studying the drawings and by a destructive test of one of his samples, Giesen became convinced that design and manufacturing flaws were causal factors in the failure of the seat pedestal. When Mann came onto the scene, he asked how much force was actually involved in the impact.
“Simple question,” Giesen said, “tough answer.”
In search of a practical method to obtain the answer, Giesen talked to a number of colleagues, who all agreed that it was a tough problem and, no, they didn’t have any suggestions to offer.
“Since I am not an expert in mechanical analysis of this sort, I sought the counsel of mechanical analysis experts I respect,” Giesen said. “From them I learned that quantifying an impact force defies traditional methods provided by handbooks and calculations in practical terms. You just can’t do it.”
Giesen happened to notice an advertisement for Algor’s Accupak/VE Mechanical Event Simulation software, which simulates motion and flexing in mechanical events, and computes stresses.
Although Giesen had extensive experience in mechanical and electromechanical design and systems engineering, he didn’t have any hands-on experience with finite element analysis and simulation. So he went to Algor’s Pittsburgh headquarters to work with an applications engineer to develop an impact model, over the course of three days.
Giesen used another Algor product, Superdraw III, to develop the model following the manufacturer’s design drawings. He incorporated information about the operator as provided by the lawyer. Where numerical values were not available, Giesen used conservative estimates and varied them over several iterations.
The seat assembly was modeled in an upright position using three-dimensional beam, plate/shell, and solid brick elements. The beam and plate/shell elements were used to represent the steel base of the seat and were defined using the material properties of steel from the software’s Material Library Manager.
The seat cushions were modeled using solid brick elements and were defined using a custom material. The density of custom material was defined so that the seat assembly would weigh a total of 120 pounds. Giesen researched common material property values of polyurethane foams at the University of Pittsburgh’s engineering library to establish a reasonable range of Young’s moduli for the cushions. The Young’s modulus of the seat cushioning was one of the variables that would be altered over a series of iterations.
Although the seat in the accident dropped 3 to 5 inches, Giesen decided to model conservatively and so assumed a 2-inch dead drop onto a stiff floor. The contact between the seat and the floor was modeled using Algor’s proprietary contact elements, which enable engineers to model how parts of a mechanism behave when they come into contact.
A representation of the operator in an upright posture— perched on the seat to reach the excavator’s controls—was then added to the seat assembly. The arms, legs, body, and head were modeled using solid brick elements with a density so that the weight totaled 288 pounds. The weight of the operator, 360 pounds, was discounted by 20 percent to account for the partial support from his feet on the pedals and his hands on the controls.
Boundary conditions fixed the hands where they would grasp the manipulator and the feet where the heels would rest on the floor. Beam and contact elements represented the knee and shoulder joints. The spine consisted of truss elements.
Although the human spine naturally has a curved shape, the spine was modeled straight to simplify the issue of impact force. If the backbone had been curved, it would have deflected in the analysis, thus absorbing much of the force, rather than calculating a total impact force as was intended. The Young’s moduli for the body parts was based on biomechanical information resources, including a telephone call from Algor’s offices to an expert in the field, and were varied over a series of iterations.
Standard Gravity Loading
The complete model was subjected to a standard gravity loading for the duration of one second, analyzed in 100 time steps. Giesen could watch the event unfold as it was processed.
The software displays the movement of the mechanism and stresses as they occur over time. Thus, Giesen was able to vary the stiffness of the seat cushions and body based on the behavior of the model.
He evaluated the results at the moment of impact in gs, for force amplification factor, which is a factor of how much a subject’s body weighs at the time of impact. Depending on the input variables, including height, some of Giesen’s models yielded an impact force of as much as 5 and 6 gs. However, Giesen’s final, optimized model, with the 2-inch fall, yielded 2.24 gs. Multiplying the estimated 288 pounds by the g force gave the operator an effective weight of 645.12 pounds at the moment of impact—a considerable load for the human spine to bear.
Giesen followed the event step by step on his computer screen and watched the force operating at points on the body in real time. “It was remarkable to see the reverberation in very high mechanical frequency rippling up and down the backbone,” he said.
There was no need for Giesen to swear in and tell a jury of his results. The report of his destructive examination and Mechanical Event Simulation results was one of many considerations in the subsequent settlement of the lawsuit.