Development of a Model of the Muscle Skeletal System using Adams. Its Application to an Ergonomic Study in Automotive Industry

Transcription

1 Copyright 2004 SAE International Development of a Model of the Muscle Skeletal System using Adams. Its Application to an Ergonomic Study in Automotive Industry G. Esteves IST- UTL C. Ferreira, A. Veloso and F. Brandão FMH - UTL ABSTRACT The aim of our study was to estimate the muscle skeletal load of human activities in the car industry, for that purpose MSC Adams Lifemodeler was used to build a mechanical model for a work place simulation were a electric screw driving was used. The data for model assembly was provided by video captured recordings, electromyography measurements and anthropometric data of workers. The task chosen for simulation was one of the more demanding, presenting a high-risk level for musculoskeletal disorders, considering the plant ergonomic department classification. The model allows the calculation of bone forces and joint moments of force. This will allow, in the future, to obtain the values of muscle skeletal action forces in wrist forearm and shoulder, caused by the action of the electric screwdriving tool. Results presented are from an ongoing study. INTRODUCTION TASK PRESENTATION This study will focus on the manually performed electric screw driving. The worker, operating under a structure that elevates the car enabling to work on the chassis region, has to elevate the tool to a height equal to its shoulder using both arms. Although it s mainly an elevation movement, the worker has to make several adjustments related to specify of the place and reaction due to the tool speed and torque action. WORKING TASK SIMULATION - The working place was recreated at the experiment and testing room of the automotive plant, and all the equipment used is the same as used on the actual workplace. The subjects were experienced workers in this operation, and the experiment-induced changes should be considered as low or inexistent as related to a possible misleading accuracy in task normal operations. The main changes were related with the cloths used and on the fact that the chassis was immobilized. The environment had a strong resemblance to the line. DATA FONTS Directly taken from the task The working region was covered with three 50 Hz video cameras that provide the 3-D data for worker kinematics, the 3D reconstruction was performed using in APAS [1]. Anthropometrics data from the workers were obtained for further use in MSC Adams Lifemodeler [2, 3]. Surface electromyography from shoulder, cervical and dorsal muscles from the right side of the body was obtained in order to evaluate the level of mechanical load produced by the relevant muscles. EMG values will update and refine the model. Used in model parameterizations GeBOD anthropometric database is used to build the model body segments configurations; this database will relate the data of age, height, weight and sex (user defined) with the size of the several ellipsoids that bring together model parts. Values of joints stiffness and damping were provided by the tutorials of MSC Adams Lifemodeler, from Biomechanics Research Group. Technical document related to tool was provided by AtlasCopco, the tool is the electric screw driver model ETV S CTADS.

2 MAIN SECTION THE MODELING PROCESS: 1. Choice of worker and task 2. Trimming and synchronization the video recording for task analysis 3. Collecting and filtering process of 3-D data 4. Model parameterization 5. Tool parameterization 6. Human environment actions parameterization CHOICE OF WORKER AND TASK - Fifteen workers were involved in the observations. For model construction the subject chosen matched medium Gebod database, being also the subject with the height and weight closest to the groups average. The work task consisted on 4 screw actions performed in 1.30 minutes one of the repetitions was selected for simulation. COLLECTING AND FILTERING PROCESS OF 3-D DATA - This sub-process allows us to obtain displacement, velocity and accelerations data of body segments from the three cameras views. For each subset of images a set of Control points coordinates was digitized, they will provide the data, in Cartesian coordinate system using the Direct Linear Transformation algorithm [4]. Workers body and tool were marked with several retroreflectors tape making simpler the location and digitizing of a total of 17 body landmarks, the full list with description is provided in Appendix 1. Figure 1- Segments Construction JOINTS - The model joints that connect all individual body segments are a combination of Hybrid III joints on the upper body for passive stabilization of the upper torso actions during direct dynamics simulation. A passive set of joints will be used in the inversedynamics. The complete set of values is provided in Appendix 1. Right Arm and Left Arm all joints (Scapular, Shoulder, Elbow and Wrist) are Passive Joints with Nominal Stiffness an Joints Stops, with the values provided by MSC Adams Lifemodeler. Lower body joints - are not subjected to the same mechanical load and they are not the core of this study; their parameterization is Hybrid III factor 1.4. For the case of hips, knee and ankle are set as Fixed. The body landmarks trajectories were smoothed with a low pass digital filter of 5 Hz. The software used was Apas Modules Digitize, Filter and Transform. MODEL PARAMETERIZATION -This sub-process is the starting point for the use of 3-D motion analysis information provided by the entire descriptions. There are several steps to follow; their description and options token during this process in order to obtain a model that can simulate reality as close as possible for further results interrogation. Segments- with the data taken from the worker, close to average of GeBod data base. LifeMode generates all body segments, the model generated is shown in Figure1. If those values where too different then ellipsoids bodies are added to fit real proportions. A matrix with the following values is generated. Age- 384 months Sex-Male Height mm Weight Kg Figure 2 - Creation of Joints POSTURE - The model is positioned into body segments configuration that fits the task; the values of joint angles are presented in appendix nº 3. This operation is needed because the motion that was recorded started at a different configuration from the default-starting configuration of LifeMod. The values were obtained from 3D kinematics has described on Figures 3 and 4.

3 From the available elements, belonging to upper and lower body, we have chosen the ones that are directly involved in the movement, in this case only upper body. Motion agents in hand, wrist and shoulder for both sides of the body and one in the top of the tool are the most important for this study. As an example in Figure 5 of model with motion agents only in the right side The values of translational stiffness and translational damping are the default values for this initial simulation process. Figure 3- Placing model, isometric view Tool parameterization - The tool model was ETV S CTADS from AtlasCopco. It was created a simulation of this tool, consisting of two rigid bodies, one cylinder and a parallelepiped box. The parallelepiped box that stands for the tool driving motor has the following measures: length 435mm, height 22mm and depth 22 mm. The extension-reaching part represented by a cylinder has the following measures: length 2mm and a radius of 8.75mm. This cylinder was designed to performed torque function that stands for the regular working conditions. These torque values may vary from 10 Newton/meter to 50 Newton/meter during the considered simulation. Figure 4- Placing the model in to position MOTION Displacement is described in terms of Cartesian coordinates, these values will be set to the dummy for the inverse dynamics simulation. The software generates motion agents that are massless parts that are fixed the body segments using spring elements. Through this attachment, motion agents are motion influencers not motion governors. This accommodates the geometric differences between the body model and the actual human subject, as well as motion agent location discrepancies due to kinematics data errors. Figure 6- Tool created in Adam View They form an angle of 90 degrees and are rigidly attached; the total weight in accordance with AtlasCopco technical data sheet is 1.9 Kg. For that purpose Aluminum basic properties where altered to the following values: Young's Modulus: (7.1705E+010 (Newton/meter**2)) newton/mm**2 Poisson's Ratio : 0.33 Density: 2.04E-006 (2040.0(kg/meter**3)) kg/mm**3 Total weigh in simulator equal to Kg Figure 5- Motion agents positioned in the right side

4 HUMAN ENVIRONMENT PARAMETERS -The model interacts with the environment in two regions, the floor and the tool. For the interaction with the floor two normal force elements between the floor and both feet were consider, the value of the normal force is given in Appendix 3. tissues used in this task are in current modeling; the major acting muscles of right arm will be added to the model so that the results taken from electromyography can update the model biofidelity. TOOL PARAMETERIZATION - One of the acting forces of this electric tool is its torque due to movement. For the purpose of having a base index of this task, joints torques and angles were calculated considering the weight of the tool only, additionally and power tool action torque was added. Important measures should arise of this fact. In the videos is obvious that not only the up and down movement of carrying the tool is important when studying the muscle skeletal charge but also some of the torque that the tool applies to the limbs. CONCLUSION Figure 7- Model in solid fill Was assumed that in all movement both hands always were firmly connected to the tool, the initial value of the bushing force was the one needed for solid interaction with the tool. For now the simulations performed show a process of modeling using state of the art software, as well as opens the possibilities to perform several major task changes, frequency, tool weight, tool torque, body segments refinements in order to fit more accurate medical data. Future results will include video files enabling the viewer the comparison between video 3-d recording and live simulation. So it will be possible to give the necessary data related to muscle skeletal load and there distribution over the different body areas allowing a quantitative and qualitative evaluation of mechanical load on the musculoskeletal that should be associated with engineering methods of human activities in the car industry. REFERENCES 1. Ariel, G., APAS - Ariel Performance Analysis System. 2003, Ariel Dynamics. 2. McGuan, S., LifeMOD. 2003, Biomechanics Research Group Inc. 3. ADAMS/View. 2003, MSC Software. 4. Kwon, Y.-H. (1988). kwon3d Motion Analysis Web. Kwon, Young-Hoo. Retrieved 1/12/2003, 2003, from the World Wide Web: Figure 8- Model with main bushing forces THE MODELING PROCESS- FUTURE STANDINGS MODEL PARAMETERIZATION - the values attributed to joints allowed taking data from the direct dynamics simulation. The process reveals that the combination presented in Appendix 1 as a functional working combination between HIII Crash test Dummy characteristics, Passive Joints values, could be used in this study and for similar activities. The values of torque and force functions recorded from previous simulation allow us to obtaining results such as reaction force in wrist, elbow and scapular region. Soft

5 APPENDIX APPENDIX 1 General parameterization used for Joints in Adams Lifemod $ JOINT_DATA [JOINT_DATA] UPPER_NECK_X ='FIXED,' UPPER_NECK_Y ='FIXED,' UPPER_NECK_Z ='FIXED,' LOWER_NECK_X ='FIXED,' LOWER_NECK_Y ='FIXED,' LOWER_NECK_Z ='FIXED,' THORACIC_X ='FIXED,' THORACIC_Y ='HIII,1.4,' THORACIC_Z ='FIXED,' LUMBAR_X ='FIXED,' LUMBAR_Y ='HIII,1.4,' LUMBAR_Z ='FIXED,' RIGHT_SCAPULAR_X ='FIXED,' RIGHT_SCAPULAR_Y ='FIXED,' RIGHT_SCAPULAR_Z ='FIXED,' RIGHT_SHOULDER_X ='PASSIVE,1000.0,10.0,90.0,- RIGHT_SHOULDER_Y ='HIII,1.4,' RIGHT_SHOULDER_Z ='PASSIVE,1000.0,10.0,45.0,- RIGHT_ELBOW_X ='PASSIVE,1000.0,10.0,1.0,-120.0,1.0E+006,' RIGHT_ELBOW_Y ='PASSIVE,1000.0,10.0,90.0,-90.0,1.0E+006,' RIGHT_ELBOW_Z ='HIII,1.4,' RIGHT_WRIST_X ='FIXED,' RIGHT_WRIST_Y ='FIXED,' RIGHT_WRIST_Z ='FIXED,' LEFT_SCAPULAR_X ='FIXED,' LEFT_SCAPULAR_Y ='FIXED,' LEFT_SCAPULAR_Z ='FIXED,' LEFT_SHOULDER_X ='PASSIVE,1000.0,10.0,90.0,- LEFT_SHOULDER_Y ='HIII,1.4,' LEFT_SHOULDER_Z ='PASSIVE,1000.0,10.0,45.0,- LEFT_ELBOW_X ='PASSIVE,1000.0,10.0,1.0,-120.0,1.0E+006,' LEFT_ELBOW_Y ='PASSIVE,1000.0,10.0,90.0,-90.0,1.0E+006,' LEFT_ELBOW_Z ='HIII,1.4,' LEFT_WRIST_X ='PASSIVE,1000.0,10.0,45.0,-45.0,1.0E+006,' LEFT_WRIST_Y ='HIII,1.4,' LEFT_WRIST_Z ='PASSIVE,1000.0,10.0,90.0,-90.0,1.0E+006,' RIGHT_HIP_X ='FIXED,' RIGHT_HIP_Y ='FIXED,' RIGHT_HIP_Z ='FIXED,' RIGHT_KNEE_X ='FIXED,' RIGHT_KNEE_Y ='FIXED,' RIGHT_KNEE_Z ='FIXED,' RIGHT_ANKLE_X ='FIXED,' RIGHT_ANKLE_Y ='FIXED,' RIGHT_ANKLE_Z ='FIXED,' LEFT_HIP_X ='HIII,1.4,' LEFT_HIP_Y ='HIII,1.4,' LEFT_HIP_Z ='HIII,1.4,' LEFT_KNEE_X ='FIXED,' LEFT_KNEE_Y ='FIXED,' LEFT_KNEE_Z ='FIXED,' LEFT_ANKLE_X ='FIXED,' LEFT_ANKLE_Y ='FIXED,' LEFT_ANKLE_Z ='FIXED,' $ POSTURE_DATA [POSTURE_DATA] Upper_Neck ='0.0,0.0,0.0,' Lower_Neck ='0.0,-20.0,0.0,' Thoracic ='0.0,-15.0,0.0,' Lumbar ='0.0,-10.0,0.0,' Right_Scapular ='0.0,0.0,0.0,' Right_Shoulder ='-10.0,-14.0,0.0,' Right_Elbow ='-120.0,0.0,0.0,' Right_Wrist ='0.0,-90.0,0.0,' Left_Scapular ='0.0,0.0,0.0,' Left_Shoulder ='0.0,-30.0,0.0,' Left_Elbow ='-105.0,0.0,30.0,' Left_Wrist ='-70.0,-90.0,-30.0,' Right_Hip ='0.0,0.0,0.0,' Right_Knee ='0.0,0.0,0.0,' Right_Ankle ='0.0,0.0,0.0,' Left_Hip ='0.0,0.0,0.0,' Left_Knee ='0.0,0.0,0.0,' Left_Ankle ='0.0,0.0,0.0,' APPENDIX 2 List of markers that were digitalized for motion analysis purpose. Lower body 1. right foot center 2. right ankle center 3. right knee center 4. right hip center 5. left hip center 6. left knee center 7. left ankle center 8. left foot center Upper body 1. right hand center 2. right wrist center 3. right elbow center 4. right shoulder center 5. left hand center 6. left wrist center 7. left elbow center 8. left shoulder center APPENDIX 3 [POSTURE_DATA] Upper_Neck =' 0.00, 0.00, 0.00,' Lower_Neck =' 0.00, 0.00, 0.00,' Thoracic =' 0.00, 0.00, 0.00,' Lumbar =' 0.00, 0.00, 0.00,' Right_Scapular =' , 0.00, ,' Right_Shoulder =' -5.00, 15.00, 5.00,' Right_Elbow =' , 15.00, 5.00,' Right_Wrist =' , , ,' Left_Scapular =' 0.00, 0.00, 0.00,' Left_Shoulder =' 15.00, 0.00, 0.00,' Left_Elbow =' , , 0.00,' Left_Wrist =' 0.00, , 30.00,' Right_Hip =' 0.00, 0.00, 0.00,' Right_Knee =' 0.00, 0.00, 0.00,' Right_Ankle =' 0.00, 0.00, 0.00,' Left_Hip =' 0.00, 0.00, 0.00,' Left_Knee =' 0.00, 0.00, 0.00,' Left_Ankle =' 0.00, 0.00, 0.00,' Values of bushing forces between right and left feet and floor Damping: 0.1 (0.1(Ne-sec/mm)), 0.1 (0.1(Ne-sec/mm)), 0.1 (0.1(Ne-sec/mm)) newton-sec/mm Stiffness: 1.0E+005, 1.0E+005, 1.0E+005 newton/mm Values of bushing forces between right and feft hand and tool Damping: 1.0E+007, 1.0E+007, 1.0E+007 newtonsec/mm Stiffness: 1.0E+007, 1.0E+007, 1.0E+007newton /mm Tdamping: 1.0E+007, 1.0E+007, 1.0E+007newton-mmsec/deg Tstiffness: 1.0E+007, 1.0E+007, 1.0E+007newton - mm/deg