音声ブラウザご使用の方向け: SKIP NAVI GOTO NAVI

Job accommodation for repetitive strain injury on the automobile assembly line

Michael J. Rosen1,2, T. Ashwin Raj1, James A. Weidhaas2 1Department of Biomedical Engineering and 2Rehabilitation Engineering Program, Department of Orthopaedic Surgery University of Tennessee, Memphis

Abstract

As part of a project supported by a contract from Saturn Corporation, an assembly line worker with lateral epicondylitis related to job tasks has been the focus of an accommodation project. Observations strongly suggested that her injury was caused in large part by use of a six-pound electric bolt driver for bolting doors to car frames. The supination and wrist flexion loads of the original job have been eliminated and reduced, respectively, by means of a "three-axis-tool-holder". This mechanical adaptation supports one end of the tool on the door-mounting fixture while allowing the operating end to move as it needs to in order to access all four bolts. Early evaluation results indicate work rates comparable to unaided tool use and a high level of worker enthusiasm following initial skepticism.

Background

In June of 1994, Saturn Corporation, in Spring Hill Tennessee, awarded a contract to the Biomedical Engineering Department (BME) and the Rehabilitation Engineering Program (UTREP) of the University of Tennessee, Memphis. The purpose of the eighteen-month project was to establish what could be done to permit assembly line workers with intractable repetitive strain injuries (rsi) to return to their jobs Ñ by applying the mix of knowledge and skills available from the two funded programs. The UT Memphis campus is known as the Health Science Center and is home to the Colleges of Medicine, Allied Health and Graduate Health Sciences. The project research team combines the clinical skills and knowledge of anatomy and physiology one would expect to find on such a campus with experience in biomedical product design. The project has been collaborative, making use of in-house knowledge at Saturn in industrial health, workplace ergonomics, manufacturing engineering and tool design. At Saturn, as at virtually all manufacturing firms whose work force is engaged in manual and power-assisted assembly, rsi is a cause of concern. Seemingly benign loads and postures, repeated many times Ñ even with carefully-planned rotations and rest periods Ñ can cause or aggravate inflammatory conditions in muscle, tendons and nerves which may become chronic and disabling for some workers. It has been estimated that 48% of recordable industrial ill health is attributable to rsi, and that the annual cost of rsi-related medical bills and lost work time is $27 billion [1]. While Saturn has taken a proactive and progressive approach to reducing the occurrence of such injuries. there is currently a group of workers who are unable to return to their original jobs because of the risk of aggravating their rsi's. The carefully cultivated autonomy of factory floor teams at Saturn and the family-like loyalty their members feel all mitigate in favor of honoring the desire of injured workers to return to their original teams. This fundamental goal set up the engineering challenge the project has attempted to meet.

Problem and Process

In the first phase of the project, the ergonomic and medical staffs at Saturn reviewed the records of their restricted workers to identify four candidates for involvement. Criteria for inclusion included a clear-cut rsi diagnosis, a willingness to cooperate and contribute ideas, an absence of complicating factors and history, and an injury to the wrist, elbow, shoulder or neck. Four potential participants were identified, one with an injury at each of those sites. The research team's task, under the terms of the contract, has been to develop accommodations for these workers which permit them to return to their original teams and perform the full rotation of tasks at acceptable speed and minimal risk of reinjury. Also, any accommodation installed on the assembly line must not impede uninjured workers, i.e. must be found attractive for general use or stay out of the way of workers who do not need to use it. The remainder of this paper documents the team's work with project participant S whose diagnosis is right lateral epicondylitis, known more commonly as "tennis elbow". The Doors Team to which she is expected to return is responsible for bolting the doors to the automobile "space frame" (the rigid structure which supports the body panels and drive train). The UT team undertook a sequence of four major tasks:

1- Identify the target assembly operation(s), i.e. those most likely to be responsible for the worker's rsi (diagnosis confirmed by UT physicians and therapists consulting to the UT research team). This phase required application of both clinical wisdom and knowledge of the musculoskeletal anatomy and mechanics of the involved body structures. Its outcome offers only a probability of correct identification since the clinical and research literature of rsi offers few guidelines which relate probability of injury to anthropometric variables, load, posture, task repetition rate, work-rest duty cycle, and the presence of other work-related and off-work stressors. Observational data was collected by means of detailed video taping on the assembly line and by capture of postures for statics modeling by using the Mannequinª anthropometric program (from Humancad¨).

2- Understand in detail the anatomical structures which are loaded by the target operation Ñ and how job-related external forces and torques map to internal stress those structures. Statics modeling was used to estimate the injurious loads at the lateral epicondyle. During the process of loading a bolt into the driver end of the tool, S and other workers typically support its six pound weight, cantilevered from the right hand, supported only by that hand, with the forearm pronated and neutral wrist extension. This requires a 39lb-in supination moment and a 21lb-in extension moment. The muscles responsible for the forces which produce these torques have their origin at the lateral epicondyle at the elbow end of the humerus. The mechanical disadvantage of these muscles, in particular relative to the 6.5in moment arm to the tool center of gravity, results in a muscular force at the lateral epicondyle of over 400lb. It was this load which the UT team sought to reduce by an order of magnitude. In addition, the dynamic "kick" of the tool when the bolt is "driven home" with a torque of 310lb-in is transmitted to ground via the wrist extensors and the bones of the upper extremity. The accommodations were meant to provide an alternative transmission path, i.e. to avoid requiring the limb to sustain the kick.

3- Design and prototyping of alternative conceptual designs for consideration by Saturn personnel. Virtually all staff members with an interest in job performance and health on the Doors team were involved in this process, in particular S, her co-workers, and the Saturn industrial physician, ergonomics experts, manufacturing engineers and tool designers. During this lengthy and iterative stage, the UT team represented ideas as drawings and physical mockups. The Human Function Lab at UTREP was set up with a '95 Saturn space frame to assist the designers, and S was brought there to simulate use of mocked up tools and to provide design input

. 4- Installation and evaluation of line-ready hardware. Prototypes considered ready for trial use were typically demonstrated off-line first to members of the Doors team, in particular their ergonomics representative. More complex solutions were set up and simulated first at Saturn's Workplace Development Center, the development laboratory where assembly line changes are prototyped and evaluated. Each of the three designs described below was subjected to a period of actual use on the assembly line by several workers. This first evaluation period served either to reject a design for reasons inherent in the design, return it to the design team for further development to remedy less important inadequacies, or accept it as ready for hand-off to Saturn tooling designers for final modifications and replication for installation on the line and long-term evaluation.

Design Outcomes

Three designs have been prototyped and evaluated by members of the Doors team on the Saturn assembly line. The first two were found flawed in fundamental ways and abandoned for this application.

1- Belt-worn tool rest. This was by far the most economical of the three accommodations, consisting of a simple trough mounted on a standard commercial tool belt. It was intended to support some or all of the weight of the tool, thereby eliminating the need for the worker to apply a supination moment when loading bolts and reducing required wrist extension forces. It was also meant to transmit the tool kick to ground via the worker's pelvis and legs rather than upper extremities. Various versions were built incorporating some differences in materials and dimensions. All permitted the tool rest to swivel about an axis normal to the belt surface to make it possible to access all four bolts while the trigger end of the tool is continuously supported on the tool rest. The outcome of trial use by several workers was a perception that donning and doffing time was not worth the functional gain. More important was the determination that workers found it necessary to bend in uncomfortable ways to permit the driver end of the tool to reach all bolts while taking advantage of the tool rest. Finally, the belt-worn tool rest introduced a potential hazard by snagging on parts of the door-mount fixture and other equipment.

2- Flexible drive shaft and remote power unit. The essential concept in this design is that the worker needn't carry the weight of most of the tool, i.e. the motor and gear train, if it is mounted on the door-mount fixture and drives the bolts via a sheathed flexible shaft (like a heavy-duty speedometer cable). Ideally, the worker would only need to support the working end of the shaft whose flexibility would allow easy access to all bolts. Further, the reaction force for tool kick would be provided by the door-mount fixture via the shaft, thereby bypassing the worker's limbs altogether. For this application, prototype implementation of this concept fell well short of the design's conceptual advantages. The standard Atlas-Copco tool was used as the fixture-mounted power unit, driving a flexible shaft through an coupling. The 5/8in gauge of the commercial flexible shaft (S.S. White Company) required for the necessary torque results in an excessively stiff shaft which is unacceptably difficult to position at all four bolt heads. The relatively short distance between the tool mount location and the four bolts further aggravated the stiffness problem by allowing a flexible shaft only one foot long. This required a tool mount which permitted the tool to pivot about two axes and translate linearly. During trial use, it was recognized that this 3-axis mount might itself be an effective accommodation.

3- Three-axis tool holder (3ath). In its delivered (fourth prototype) form, the 3ath consists of a clamp for the standard Atlas-Copco tool mounted on a custom-fabricated universal joint to permit angular travel vertically and laterally relative to the car. This assembly is mounted on a carriage which rides on a splined shaft via a recirculating-ball bushing. The functional outcome is that the worker need only be concerned with the driver end of the tool, to which the trigger has been transferred by means of a simple push-rod linkage. Since the tool clamp supports the tool at a site two inches in from one end, part of the tool's weight counterbalances the rest. The result is that the downward load in the worker's hand is less than half the tool weight, i.e. 2lb 12oz. The entire unit is mounted on a base plate which is secured to the door-mount fixture via two quick-release clamps so that in the event of failure it can be swapped with a replacement without stopping the assembly line. The only modification to the tool itself is the use of right-angle power cord to avoid snagging on parts of the door-mount fixture. All components except commercial bearings and the splined shaft are machined from Aluminum and the device is light enough to be easily hand carried when being transported off-line.

Evaluation

At this writing, the delivered version of the 3ath is about to be used on the line for a sequence of several shifts at a time. The Saturn manufacturing engineers and tool designers are making close observations of the accommodation and its use prior to making a final decision to replicate it for the three other door-bolting stations. Their intent is to modify details of the delivered design in any way necessary to make it consistent with Saturn standards for tooling, in particular to insure sufficient reliability under assembly line conditions and speed of swapping with a replacement. An unanticipated observation from more than one member of the Saturn staff is that the 3ath will make it much more difficult to inadvertently cross-thread a bolt by driving it at an angle. The most direct evaluation of it's success will occur, of course, if S returns to the Doors team and determines whether she can rotate through the modified bolting operation with her team mates according to a regular schedule without sustaining an recurrence of her epicondylitis symptoms. In it's first on-line evaluation, which lasted over two hours, over ten workers used the 3ath to assemble left front doors on production automobiles. Worker speed with the unfamiliar accommodation was at all times adequate to maintain line speed. No mechanical difficulties were encountered and dis/mounting was accomplished during ten-minute break periods. While new users' initial attitudes ranged from enthusiasm to skepticism, expressions of approval were virtually unanimous following use. Five workers were timed to record the total time required to drive all the bolts for each door. Two were on average faster than their unaided rear-door counterpart, and three were slower. All were within the range of inter-worker and intra-worker variability typically observed with the unmodified tool. In addition to delivery of a mechanically successful job accommodation which appeals to S's co-workers, this project has had generalizable outcomes at a different level. Lessons have been learned about interactions between an academic r&d team and the workers for whom technology is being developed:

1- Timeliness is next to godliness when it comes to adhering to schedules for demonstration and delivery of prototypes. Expectations of assembly line workers, based presumably on past experience, are such that many are likely to interpret a delay of weeks in meeting a deadline as evidence that the academic team is not really committed to making a difference on the factory floor. Although any communication helps, working hardware Ñ on site Ñ speaks louder that any number of e-mail messages and phone calls.

2- Credibility capital can also be wasted by delivery of prototypes which fail to perform well. While there is an understanding on the part of people on the line that new hardware needs to be debugged, there is detectable damage to the design team's reputation each time a device iteration shows up flawed by failure to foresee an "obvious" problem.

3- The value of learning from the intended users of a new device is an often-stated lesson for new design engineers. Nevertheless, it cannot be overstated and turns out to be true in some unexpected ways. The UT team found that face-to-face conversation with workers had value not only in providing answers to questions important to the design process, but also for detecting dimensions of the task which simply hadn't occurred to them. For example, the workers' desire to have all components of the position and orientation of a tool under their control in order to finesse anomalous occurrences without falling behind line speed was an unexpected finding which helped to explain their initial skepticism about the 3ath.

4- Another value of frequent face-to-face communication between the academic design team and the workers is purely human. Being there, frequently and regularly, has the palpable effect of making one part of their family. This familiarity earns a level of cooperation with and appreciation of the project without which it is unlikely to succeed.

Acknowledgments

The authors gratefully acknowledge the continuous involvement and contributions of Mike Schlacter at Saturn who has been our link to people, facilities and knowledge in the plant.

Reference

Goldoftas, B. 1991, Hands that hurt: Repetitive motion injuries on the job, Technology Review, 43-50

-------------------------------------------------

Michael J. Rosen Rehabilitation Engineering Program The University of Tennessee, Memphis 682 Court Ave. Memphis, TN 38163 (901) 448-6448 FAX: (901) 448-7387 mrosen@utmem1.utmem.edu