line balancing considering ergonomics.pdf

Faculty of Technical Sciences Assembly Line Balancing Problem Considering Ergonomic Factors

Graduation Project Course: System Management

Skopje – October, 2011

1

Faculty of Technical Sciences Assembly Line Balancing Problem Considering Ergonomic Factors

Graduation Project Course: System Management

Mentor: Prof. Aleksandra Porjazoska Kujundziski Performed by: Nihal Kemer, 07.08/3 e-mail:[email protected]

Skopje, October, 2011

2

CONTENT.........................................................................................................................................................3 1. ASSEMBLY LINES..............................................................................................................................10 1.1. BASIC PRINCIPLES OF ASSEMBLY LINE...................................................................................12 1.2. A CLASSIFICATION OF AL..............................................................................................................14 1.2.1. ACCORDING TO THE NUMBER OF PRODUCTS OR MODELS...........................................14 1.2.2. ACCORDING TO THE TASK DURATION................................................................................15 1.2.3. ACCORDING TO THE LINE SHAPE OR LAYOUT..................................................................15 1.2.4. ACCORDING TO THE WORKPIECES FLOW............................................................................17 1.2.5. ACCORDING TO THE LEVEL OF AUTOMATION...................................................................19 1.3.1. A TYPICAL ASSEMBLY LINE......................................................................................................20 2. ERGONOMIC APPROACH...................................................................................................................22 2.1. EFFECTIVE ERGONOMICS.............................................................................................................23 2.2. GOALS OF ERGONOMISC...............................................................................................................25 2.2.1. CONTINIUOUS IMPROVEMENTS AND BREAKTHROUGHS................................................25 2.3. WORK RELATED MUSCOSKELETAL DISORDERS..................................................................26 2.3.1. STANDING AND WALKING POSTURE DURING ASSEMBLY OPERATIONS..................30 2.3.2. UPPER EXTREMITY ASSESMENT TOOLS...............................................................................32 2.3.3. HAND VIBATION EXPOSURE ON AN ASSEMBLY LINE......................................................33 3. THE RELATION OF TASK ASSIGMENT TO EXERTION FREQUENCY, DUTY CYCLE, NORMALIZED PEAK FORCE, VIBRATION ACCELEATION AND VIBRATION DURATION...................................................................................................................................................39 3.1. DETERMINATION OF HAND ACTIVITY AND PEAK FORCE MEASURES BY TASK ASSIGMENT.................................................................................................................................................39 3.2. DETERMNATION OF HAND-ARM VIBRATION MEASURES BY TASK ASSIGMENT.......41 4. OPTIMIZATION MODEL...................................................................................................................42 5. NUMERICAL EXPERIMENTS AND DISCUSSION........................................................................48 5.1. MANUFACTURING TASK DESCRITION AND PARAMETER ESTIMATION.....................48 5.2. RESULTS FROM DIFFERENT COMBINATION OF ERGONOMIC CONSTRAINTS..........51 6. CONCLUSIONS...................................................................................................................................56 APPENDIX A..............................................................................................................................................58 APPENDIX B..............................................................................................................................................60 REFERENCES...........................................................................................................................................65

3

LIST OF FIGURES FIGURE 1.1: FORD‘S CAR ASSEMBLY LINE …………………………………………….....10 FIGURE 1.2: PRECEDENCE GRAPH …………………………………………………….....12 FIGURE 1.3: SINGLE-MODEL LINE ……………………………………………………......13 FIGURE 1.4: MIXED MODEL LINE ………………………………………………...……...14 FIGURE 1.5: MULTİ-MODEL LİNE ........................................................................................ 14 FIGURE 1.6: SERIAL LINE (AN ASSEMBLY LINE OF VCR UNITS OF SONY) ………………..15 FIGURE 1.7: TWO-SIDED LINES (THE ASSEMBLY LINE OF THE TOYOTA LEXUS, CANADA)… 15 FIGURE 1.8: PARALLEL LINES ………………………………………………………..16 FIGURE 1.9: U-SHAPE LINES ………………………………………………………………..16 FIGURE1.10: CLOSED LINE ………………………………………………………………..17 FIGURE 1.11: SYNCHRONOUS LINE ………………………………………………………..17 FIGURE 1.12: ASYNCHRONOUS LINE ……………………………………………………... 17 FIGURE 1.13: FEEDER LINES ………………………………………………………………18 FIGURE. 1.14: MANUAL LINE ……………………………………………………………... 18 FIGURE 1.15: ROBOTIC LINE ………………………………………………………17 FIGURE 1.16 THE TASK DESCRIPTION AND PRECEDENCE OF ASSEMBLY TASKS WITH VIBRATION PROFILES ………………………………………………………………………………19 FIGURE 2.1: CONCEPTUAL MODEL OF THE DEVELOPMENT OF WORK-RELATED MUSCOSKELETAL DISORDERS ………………………………………………………27 FIGURE 2.2: COST ANALYSIS BY OCCUPATIONAL INJURY AND ILLNESS TYPES IN CANADA ………………………………………………………………………………27 FIGURE 2.3: FORMER BATCH WISE PRODUCTION(LEFT) AND ONE-PIECE FLOW PRODUCTION (RIHT) IN FINAL ASSEMBLY …………………………………….29 FIGURE 2.4: HAND ACTIVITY LEVEL(0- 2)CAN BE RATED USING THE GUIDELINE………………………………………………………………………………32 FIGURE 2.5: A REPRESENTATIVE WORKSTATION …………………………….34 FIGURE 2.6: CENTER FREQUENCY OF THIRD OCTAVE BANDS ……………36 FIGURE 2.7: THE STRUCTURE OF THE RESEARCH ……………………………38 FIGURE 3.1: THE TLVS BASED ON HAL AND NPF. …………………………………… 40 FIGURE 5.1: THE TASK DESCRIPTION AND PRECEDENCE OF ASSEMBLY TASKS WITH VIBRATION PROFILES ……………………………………………………48 FIGURE 5.2: THE ACCELERATION AND THE COORDINATE FIGURES OF TWO SCREWDRIVERS …………………………………………………………………..50 FIGURE 5.3 RESULTS FROM DIFFERENT CONSTRAINT COMBINATIONS ………….52 FIGURE 5.4: HAL TLV OF LBMC, LBMCH, LBMCV AND LBMCHV …………52 FIGURE A.1:ONE-THIRD OCTAVE BAND SPECTRA 3,4 CM DIAMATERAUTOMATİC SHUT-OFF SCREDRIVER ………………………………………………………..58 Figure A.2:ONE-THIRD OCTAVE BAND SPECTRA FOR THE 2,5 DIAMETER CLUTCH SCREWDRIVER IN THE SLIPPAGE OF THE CLUTCH CONDITION…………………………………………………………………………58

4

LIST OF TABLES TABLE 2.1: FORMER AND NEW SITUATIONS IN FINAL ASSEMBLY AND PRINT ASSEMBLY AND THE MAIN NEW ELEMNTS............................................................... 29 TABLE 2.2: HAND ACTIVITY LEVEL(0-10) IS RATED BASED ON EXERTION FREQUENCY AND DUTY CYCLE................................................................................... 32 TABLE 2.3: MEAN AND EIGHT HPUR VIBRATION EXPOSURE DURING RUN AND CLUTCH CYCLES FOR POSITION REQUIRING OPERATION OF POWER SCREWDRIVERS ........................................... 35 TABLE 3.1: TLVS FOR EXPOSURE OF THE HAND TO VIBRATION IN EITHER X, Y, Z AXIS .................................................................................................................................. 42 TABLE 4.1: REFERENCE TABLE OF HAL WITH M = 5 .................................................... 45 TABLE 4.2: THE SET OF RANGES OF EXERTION FREQUENCIES ................................ 46 TABLE 4.3: THE SET OF RANGES OD DUTY CYDES ...................................................... 46 TABLE 4.4: THE CONVERSION OF INTEREDIATE VALUE TO HAL ....................... 46 TABLE 4.5: THE CONVERSION OF DAILY VINRATION TO ACCELERATION TLV .... 47 TABLE 5.1: DATA EXERTION NUMBER, CYCLE TIME AND NPF................................. 49 TABLE 5.2: MALE POWER GRIP STRENGTHS IN NEWTON(N)..................................... 50 TABLE 5.3: ACRONYMS FOR THE LINE BALANCING MODELS .................................. 51 TABLE 5.4: RELATIVE VALUE OF ERGONOMIC MEASURES COMPARED TO LBMC CASE .................................................................................................................................. 54 TABLE 5.5: VIBRATION RESULTS ................................................................................... 55 TABLE B.1 HAND ACTIVITYRESULT OF LBM ............................................................... 60 TABLE B.2 HAND ACTIVITYRESULT OF LBMCH.......................................................... 61 TABLE B.3 HAND ACTIVITYRESULT OF LBMCV.......................................................... 62 TABLE B.4 HAND ACTIVITYRESULT OF LBMCHV ....................................................... 63 Table B.5 DETAIL VIBRATION RESULTS…………………………………………....... 64

5

NOMENCLATURE The mathematical symbols used in this study are introduced as follows.

Indexes

f = The one-third octave band number

h = Hand number; h=1, 2

i, j = Manufacturing task

o = The direction of vibration, o=X, Y, Z

p = The range number of

q = The range number of

s = Worker or station; s=1,2,….,M

u = One dimensional cell value fort he table of hand activity level

Parameters

CN = The maximum Colomn Number

CT = The cycle time of the assembly line

= The HAL value corresponding to u

K = Weighting factors for acceleration calculation

NC = The number of cycles finished per day which equals to WT/CT 6

RN = The maximum row number

WT = The total working time per day

= The f-th frequency acceleration

= The frequency -weighted,rms acceleration of task i,directon o

= The frequency-weighted acceleration

= 5% up force time of one task of hand h

= The number of exertions of hand h during one task

= The processing time of task i = The duraction of vibration of task i

Variables

1(

= The indicator variable;one if

1(

) = The indicator variable;one if

belongs to

, zero otherwise

belongs to

, zero otherwise

= The total daily vibration exposure duration of at station s

= The equivalent,frenquency-weighted component acceleration of station s,direction o

= Column number in the table of duty cycle

= duty cycle of workers s

= Exertion frequency of worker s 7

= Hand activity level of worker s, hand h = The limit of the equivalent,frequency –weighted component acceleration of worker s, direction o

= Peak hand force of worker s, hand h = Peak hand force of worker s, hand h

= Row number in the table of exertion frequency

= The intermadiate variables for calculating HAL;

=

+RN. (

= The binary variable of task assigment fort ask i, station s

= The binary variable of hand activity level of each hand of each worker in each station

z = Objective function value

8

ABSTRACT Musculoskeletal disorders (MSDs) related to the workplace are among the most costly problems in today‘s society. The low-back, neck, shoulders, and upper limbs are the body parts most subject to risk. Among the factors associated with the risk for developing workrelated MSDs are individual, physical workplace, organizational, and psychosocial factors. After the ergonomic intervention the sick leave due to low back problems decreases, so that the company benefits economically from such sick leave reduction because of lower replacement costs, learning costs, and productivity rates of new workers. Also, the properly planned and implemented ergonomics result in significant economic benefit due to productivity and quality increases. The conventional approaches to decreasing work related musculoskeletal disorders (WMSD) risks in the assembly lines include slowing the work-pace or applying job rotations. These adjustments usually focus on individual assembly workers at the station level but not the work allocation among the workers at the whole assembly line level, and thus may decrease the line productivity. To avoid these negative effects, some research started considering ergonomic characteristics at the line level, such as balancing ergonomic burdens by proper work assignment among workers. Thus, the aim of this diploma work was to make a survey of a literature which integrates ergonomic factors in resolving assembly line balancing problems, especially considering ergonomic measures for upper extremity musculoskeletal disorders. A suitable model can effectively control the exposure levels in the upper extremity by proper work assignment, compared to the conventional approaches, and does not decrease production rates considerably. The linear models allowing ergonomic and productivity measures to be integrated as a mixed-integer programming model for assembly line design, are especially attractive. Linear models are developed to link work-worker assignment to the measures of hand activity and hand-arm vibration. As productivity measures, conventional assembly line design criteria are considered, such as cycle time and the number of workers. In addition, these linearization methods can be generalized in order to incorporate ergonomic measures in tabulated forms into assembly line design problems. This literature survey will act as a baseline for our future research considering similar ergonomics problems.

9

1. ASSEMBLY LINES An assembly line is a flow-oriented production system where the productive units performing the operations, referred to as workstations, are aligned in a serial manner. The workpieces visitstations successively as they are moved along the line usually by some kind of transportationsystem, e.g. a conveyor belt. Each workstation repeatedly performs a set of tasks in order to produced or manufacture a specific product. Tasks require certain time to be processed and are related among one another according to the existing technological constraints. Originally, assembly lines were developed for a cost efficient mass-production of standardized products, designed to exploit a high specialization of labour and the associated learning effects. The most famous example of an assembly line is the production plant of Henry Ford (Figure 1.1), and the main focus of ALBP is how to distribute the entire workload to the work stations of an assembly line [1].

.

FIGURE1.1: Ford‘s car assembly line T-model components were manufactured in the first moving line using the ideas of work division to decrease the production cost per unit and to allow massive production. However, the work division ideas and this kind of configurations date from much earlier times. The Venetian Arsenal (considered the world first factory) for instance, developed methods of 10

mass-producing warships which were much faster and required less wood. At the peak of its efficiency in the early 16th century, the Arsenal was able to produce nearly one ship per day on a production-line basis not seen again until the Industrial Revolution. In 1799, Eli Whitney introduced the assembly lines in the American manufacturing system. In 1901 Ransom Eli Olds patented the first assembly line concept and his Olds Motor Vehicle Company was the first factory in America to mass-produce automobiles. Was until 1913 when Henry Ford perfected the assembly line concept; nowadays, the first assembly line for building cars is attributed to him [2]. Since the times of Henry Ford and thefamousmodel-T,however, product requirements and thereby the requirements of production systems have changed dramatically. Although, assembly lines are most commonly found in the automotive industry, many other sectors are also organized in assembly lines. This is the case for most daily life goods, as, for example, the final assembly of electrical products such as coffee machines, washing machines, refrigerators, radio, TV and personal computers. More recently, assembly lines have gained importance in low volume production of customized products as well as in service systems [2].

Assembly lines have many advantages and disadvantages: The benefits that assembly lines provided for businesses are mainly the following: -

-

Provides a regular flow of material; allows using of power looms and human capacity at the highest level; aims to minimize blank of duration; distributes empty times between work stations properly; minimizes production costs. In addition, in order to respond to diversified costumer needs, companies have to allow for an individualisation of their products. For example, German car manufacturer BMW offers a catalogue of optional features which, theoretically, results in1032 different models. Multi purpose machines with automated tool swaps allow for facultative production sequences of varying models at negligible setup costs. This makes efficient flow-line systems available for low volume assembly-to-order production and enables modern production strategies like mass-customization, which in turn ensures that the thorough planning and implementation of assembly systems will remain of high practical relevance in the foreseeable future.

Among the disadvantages one can account: -

There are low-skilled workers on assembly lines and employees are doing the same jobs consistently because of this there is monotony. In addition, rates of change in demand is directly linked to the efficiency of the production system. Also, due to the high level of automation, assembly systems are associated with considerable investment costs. Therefore, the (re)-configuration ofan assembly line is of critical importance for implementing a cost efficient production system. 11

-

Configuration planning generally comprises all tasks and decisions which are related toequipping and aligning the productive units for a given production process, before the actual assembly process can start. This includes setting the system capacity (cycle time, number of stations, station equipment) as well as assigning the work content to productive units (task assignment, sequence of operations).

1.1.Basic principles of assembly line

Processing tasks: a processing task i (task, here after) is an indivisible working unit which has associated processing time t i. The total work required to manufacture a product in an assembly line is divided into a set of n tasks. Workstations: are the line component where tasks are processed, and can involve a human or robotic operator, certain equipment and some specialized processing mechanisms. Cycle time ct: is the time available in each workstation to complete the tasks required to process a unit of product -the production rate is equal to 1/ ct units of product per time unit. The cycle time is also defined as the time interval between the processing of two consecutive units. Precedence relations: are defined by the technological precedence requirements that determine the partial order in which tasks can be performed in the assembly line. A task cannot be processed until all its immediate predecessors have already been processed. Precedence relations are normally represented by a precedence diagram. Figure 1.2 shows an example precedence graph with n=9 tasks having task times between 2 and 9 (time units) [1].

FIGURE 1.2: Precedence graph

Workstation load Sj: is the subset of tasks assigned to workstation j.

12

Workstation time t(Sj): is the sum of the times ti of all tasks assigned to workstation j.

Workstation idle time It j: is the difference between the cycle time and the workstation load.

Line balancing: is the process of distributing the n tasks among the m workstations in such a way that precedence constraints and other constraints are satisfied; aiming at optimizing a given efficiency measure. Classical objectives seek to minimize m for a desired cycle time ct, or to minimize ct given m. There exists a great variety of configurations involving assembly lines, which are characterized according to diverse criteria. Amongst others, these include the layout and shape of the line, the number of products and models being processed in the line, types of workstation and the variability of the task processing times [1].

1.2. A classification of assembly lines (AL) A classification of assembly lines is usually done based on the number of products or models produced, tasks durations, shape or layout of the line, the flow of the workpieces and the level of automation of the line.

1.2.1. According to the number of products or models

Single-model line: is the classical configuration in which a single model of a unique product type is produced (Figure 1.3).

FIGURE 1.3: Single-model line

13

Mixed-model line: several variants of a basic product, referred to as models, are produced simultaneously in the line (see Figure 1.4). The production process does not involve setup times since all models require basically the same manufacturing tasks. Units of different models are produced in a mixed sequence [2].

FIGURE 1.4: Mixed model line Multi-model line: different models with significant differences amongst one another are processed in the line. Therefore, sequences of batches are processed, containing either the same model or a group of similar models, involving intermediate setup tasks (Figure 1.5) [2].

FIGURE 1. 5: Multi-model line

1.2.2. According to task duration

Deterministic: all task processing times are fixed and known with certainty. Stochastic line: task processing times may be significantly affected from different sources of variability such as, for example, the ability or motivation of human operators. Therefore, the processing time of one or more tasks is considered to be probabilistic [2]. Dependent line: tasks processing times are not fixed but dependent, for example, on the type of workstation to which the task is assigned, on theoperator or on the processing sequence. Dynamic line: processing times vary over time and can be reduced in successive cycles due to improvements in the assembly process or due to learning effects (for example, when operators become familiar with the tasks [2].

14

1.2.3. According to the line shape or layout

Serial lines: products units are processed throughout a group of workstations that are consecutively arranged in a straight line such as, for example, a conveyor belt (Figure 1.5)

FIGURE 1.6: Serial line (An assembly line of VCR units of Sony)

Two-sided lines: consist of two serial lines in parallel, in which pairs of opposite workstations (left-hand side and right-hand side) process simultaneously the same workpiece. This configuration is commonly found in the automotive industry (Figure 1.6). Some tasks can be assigned only to one side (e.g. mount the left car wheel), some tasks can be assigned to either side (e.g. install the hood ornament), and some tasks must be assigned to both sides of the line simultaneously (e.g. install the rear seat) [2].

FIGURE 1.7: Two-sided lines (The assembly line of the Toyota Lexus, Canada)

Parallel workstations: in this case two o more workstations are put in parallel; hence, the work pieces can be distributed between several workstations that perform an identical set of tasks.

15

Parallel lines: this type of configuration can be considered when the production system involves multiple products, in which each line is designed for one product or family of similar products (Figure 1.8) [2].

FIGURE 1.8: Parallel lines

U-Shaped lines: the workstations are arranges in a U-shaped line. Both tops of the line are closed to each other forming a U (Figure 1.9). The workstations may work during the same cycle on two or more workpieces at different positions on the line.

FIGURE 1.9: U-shape lines

Circle/closed lines: in this type of lines the workstations are arranged around a circular conveyor belt (or similar mechanism), as can be seen in Figure 1.10. A workpiece moves around being processed as it visits the workstations, until the last task have been performed. [2].

16

FIGURE 1.10: Closed line

1.2.4. According to the workpieces flow

Synchronous lines: in synchronous lines, also referred to as paced lines, all workstations have a common cycle time. Therefore, all workstations start processing at the same time and advance the workpieces simultaneously. Synchronous lines have a fixed production rate (Figure 1.11).

FIGURE 1.11: Synchronous line

Asynchronous lines: in these lines all workstations can work at different speeds; thus, workpieces are transferred whenever the required tasks are completed. The workstations are linked by buffers to store the workpieces that cannot advance to the next workstation due to it is processing another workpiece (Figure 1.12).

FIGURE 1.12: Asynchronous line

Feeder lines: feeder lines are supplementary lines that provide a main line with subassemblies. Figure 1.12 shows an example of an assembly process of an aeroplane, which consists of four lines feeding the main line [2]. 17

FIGURE 1.13: Feeder lines

1.2.5. According to the level of automation

Manual lines: in manual lines the tasks are performed by human operators. These lines are common when workpieces are fragile or are of special importance. Harley Davidson‘s motorcycles, for example, are 100% assembled by hand as shown in Figure 1.14.

FIGURE. 1.14: Manual line

FIGURE 1.15: Robotic line (http://encarta.msn.com/media ˙701765960/Robot˙Assembly˙Line.html (visited on February 2004)

The characterization of line, to a great extent, determines the type of balancing problem that is to be solved. For example a single-model, serial, paced, deterministic line merely entails the aassigment of tasks to the workstations -the simplest balancing problem. However,for 18

other line configurations the balancing problem comes toget her with a problem of sequencing the models, wheras a multi-model line also implies a lot sizing problem. Paralel lines involve a decision problem concerning the number of lines that needs to be installed. Robotic lines (Figure 1.15), on the other hand, involve the assignment of both tasks and robots to the workstations. Asynchronous lines requires of the positioning and dimensioning of buffers: -

vand whenever feeder lines are considered, the production rates of the available lines have to be synchronized [2].

It is evident that industrial systems involve a great variety of characteristics and problem variations. However, due to their complexity, most literature studies on production and manufacturing have addressed problems which do not consider many of the requirements and constraints present in real systems. However, in the last years a considerable effort has been done towards filling the gap between problems addressed in research works and real-world applications [2].

1.3. Assembly Line Balancing Problems (ALBP) The distribution of the entire workload to the work stations of an assembly line, as the main focus of assembly line balancing problems (ALBP) have been extensively studied [1,3-6]. The assembly line balancing problems, firstly proposed by Helgeson originated with the invention of the assembly line. Figure 1.15 shows a typical assembly line.

FIGURE 1.16 Typical assembly line

19

1.3.1. A typical assembly line

As previously mentioned, the Assembly Line Balancing Problem (ALBP) consists in assigning a set of indivisible tasks to an ordered sequence of workstations in such a way that precedence constraints are maintained, the workload of each workstation does not exceed the cycle time and a given efficiency measure is optimized. The term balancing arises from the fact that the workload of each workstation is to be balanced. Since a perfect balance (i.e., an identical load for all workstations) is rarely achieved, workstations idled time becomes a main optimization objective. On the other hand, as the assembly line global cost is influenced by the number of workstations, the classical objective of assembly line balancing problem is minimize the number of workstations for a given cycle time, which is referred to as timeoriented line balancing. Furthermore, production rate can be maximized by minimizing the cycle time of a given number of workstations. Problems that seek to minimize costs are regarded to as cost-oriented line balancing. On the other hand, profit oriented are those which implicitly consider the profit attained by the line. The objectives of ALBP are to minimize cycle time, reduce number of workstations, or level manufacturing workload. Often, more that one efficiency measure is to be optimized. The following are some of the minimization criteria that have been considered in literature studies: throughput time (i e., the time interval between lunching a workpiece and finishing thr finished product from the line), cost of machinery and tools, inventory cost, dead time, number of buffers, line stoppage time, and variances in workstation times. Some maximization objectives include production rate (which is equivalent to minimize the cycle time), line efficiency and profit. A further objective considers that the workload of each workstation needs to be as similar as possible [4].

1.3.2. Procedures to Solve Assembly Line Balancing Problems During the first forty years of the assembly line‘s existence, only trial-and-error methods were used to solve the assembly line balancing problems. However, there have been numerous methods, both, exact and approximate, developed to solve the different forms of the ALBP. The first mathematical formulation was firstly established by Salveson. Gutjahr and Nemhauser showed that the ALP problem falls into the class of NP-hard combinatorial optimization problems. This means that although guaranteeing an optimum solution, exact methods have a problem size limitation, measured in terms of computing time; therefore, they can only be applied to problem instances with small or medium number of assembly tasks. Approximate methods (i.e., heuristics and metaheuristics) have been developed in order to overcome such a limitation, and aiming at providing good solutions that are as near as possible of the optimal solution [4]. The exact procedures applied so far are [2]:

20

-

-

-

Mixed (integer) linear programming – which is successfully used in description of the assembly line balancing problems, but it is not efficient in considering the real-world scaled problems. Dynamic Programming (DP) procedures basically transform the problem into a multistage decision process by breaking it into smaller sub-problems, which in turn are solved recursively; then the optimal solution of the sub-problems are used to construct the optimal solution of the original problem. The main drawback of these procedures is their large memory requirements. Branch – and – bound (B&B) technique - which finds the optimal solution by exploring subsets of feasible solutions. Sub-regions are formed by branching the solution space. A bounding process is recursively used to find lower or upper bounds of the optimal solution in each sub-regions.

Two main types of approximate methods are: -

Heuristic methods– bases on logic and common sense, where, at each step of the procedure, one element of the solution is chosen according to a given criteria, for example, positional weight (the summation of the task time and the processing times of all its successors), maximum task time, maximum total number of successors, minimum earliest and latest workstation and minimum slack, etc., until a complete solution is obtained. The simplest method randomly generates solutions, evaluates each one of them and keeps the best of all solutions obtained. The main drawback of classical heuristic methodsis a failure in a local optimum.

Metaheuristics methods- have been developed to overcome the previous limitations. These procedures are developed to find an initial solution and local search algorithms to move to an improved neighbor solution. In contrast to local search approaches, metaheuristics method do not stop when no improving neighbor solutions can be found. They allow movement to worsening solutions in order to avoid premature convergence to a local optimum solution. During the past decades numerous optimal approaches have been developed to solve ALBP with different characteristics, including parallel, U-type, mixed-model, two-sided, etc. Anyway, the assembly line consists of series of work stations, in which particular operations (set of tasks) are executed repeatedly. The objective of ALBP is toassigns the set of tasks to successive workstations in order to minimize the number of workstations needed for a given cycle time or maximizing the productivity [6]. Task assignment is one of the most decisions in assembly line balancing [7]. However, assembly work mainly involves activities such as fitting, adjusting, handling and quality control. The physical stresses thereby arising in automobile assembly are caused mainly by unfavourable postures, application of forces, manual materials handling and repetitive tasks. These are complemented by stresses on the sensory organs and nerves during performance of test procedures. Vehicle geometry is responsible for forced working postures in the form of, for example, (lateral) bending and twisting of the trunk and extension of the arms. Other factors can include the need to apply 21

strong forces, unfavourable load-handling situations and extreme joint angles, in some cases aggravated by heavy stress on the finger-hand-arm system from application of strong action forces and repetitive movements. Consequently, the stress accumulations occurring in assembly work are muscular (and cardiovascular), accompanied by biomechanical stresses on the spinal column and unilateral dynamic stresses on the hand-arm system. It seems possible that the physical stresses are at least a contributory cause of diseases and that these could consequently be classified as ‗work-related diseases‘. Manual materials handling and application of physical force in association with unfavourable body postures are the principal sources of stress in physical work, involving an increased health risk and possibly higher absence times from work. The potential causes of this include stresses caused by work tasks, physical and chemical environmental factors and organizational and social working conditions, for example: – workplace design (e.g. space and movement-related factors, information flow and safety aspects), – the work process in the broadest sense (i.e. everything included in the definitions of microand macro-logistics in both the immediate workplace environment and the whole building or plant), – stresses caused by noise, climatic conditions, lighting, dampness, drafts, mechanical vibrations and noxious substances, – degree of freedom in planning work actions, responsibility etc., and also cycle times, shift times and shift plans, – work climate, – remuneration system and other job-related financial factors [8].

2. ERGONOMIC APPROACH 2.1. Effective ergonomics Ergonomics is an applied engineering discipline used to design the work environment to reduce injury and illness and support human performance. A successful ergonomics process can be defined as one that is sustainable, business driven, and injury-reducing. Thus, the three critical elements of an effective ergonomics process have been identified are:  Risk reduction strategies: A proactive approach to ergonomics means that factors known to contribute to injury or illness are identified and addressed early.

22

Observation is the most basic form of assessment, which by the aid of assessment tools, firstly including the identification of risk factors, and then implementation of the solutions. The goal is to capture existing insight and information from the work force, and then quickly move to resolving problems. To take advantage of this simple approach to risk assessment, companies provide basic ergonomics awareness training and develop a process for employees to identify known challenges and provide suggestions. Providing a simple mechanism (like a feedback form) can ensure that all input is captured and reviewed [9]. So, the three basic steps of risk management are: recognition, evaluation, and control (REC). REC is a systematic approach that must involve a cycle of measuring, improving, and re-measuring to maintain a safe environment. The REC approach is simple and effective: identify the challenge through basic awareness, evaluate the degree of risk, and then implement control strategies. Those ergonomic risks that are not easily resolved require analysis with a higher level of knowledge, usually from a specifically trained group of employees, engineers, or facility managers. The assessment tools must provide quantitative data for identifying and prioritizing ergonomic risks, and they should identify alternative job improvements. They must be simple enough to be used reliably by health and safety professionals, engineers, and ergonomics committee members, but they should provide enough information to be useful. Efficient assessment tools help users to obtain information needed to make decisions with as little effort as possible. A structured risk factor survey or office needs questionnaire is usually the best approach. Some assessments make pre-approved, easy-to-implement solutions available, while others may require more advanced control strategies and an in-depth analysis to reduce risk [9]. 

The element ―fix once, repeat many‖ or FORM, is applied in avoiding the design amnesia. Namely, many companies are focused just in fixing the existing ergonomics problems, ignoring the future. This is known as design amnesia, a vicious circle in which known problems are replicated in new designs even after effective solutions to the problems have been defined and implemented. Since the designers have a greater understanding of technical requirements and not in people requirements, the design amnesia often occurs. Exclusion of the design amnesia is possible by the application of two simple mechanisms: education for the specialists to recognize ergonomics problems and easy access to the effective design solutions. With the application of FORM strategy, products, manufacturing equipment, and office workstations optimize human performance, and one can prevent design mistakes from the start. Design requires informed decision making to formulate the best solution while balancing a wide range of trade-offs. The ergonomic quality of the design is often overlooked for those who must use the product or equipment. This is simply due to the lack of education for technical personnel in the principles of ergonomics and human performance [9]. 23



Management of human capital is the third element of effective ergonomics. This element quantifies the bottom-line impact of improvements based on a healthy workforce and improved productivity. With a good design, people perform more efficiently and reliably and are less likely to get hurt. Too many business managers, however, the benefits of good ergonomics are not so clearly defined. Active support from senior management is critical to the success of an ergonomics program, so the managers should consider ergonomics as: - A good investment – the investment in ergonomics is justified by the reduction of the workers‘ compensation cost. Consider ergonomics exactly like any other important business function [9].

2.2. Goals of ergonomic Specifically, some of the goals of ergonomics are: 

  

Applied in the equipment design, as for the workplace, intend to maximize productivity by reducing operator fatigue and discomfort. Benefits provided by ergonomics application in assembly systems design are first of all linked to the reduction in occupational injury risks and in the improvement of health condition and motivation of human resources. To accommodate differences in strength and body size among different workers. To remove barriers to productivity and performance. Continuous improvements and breakthroughs.

The reduction of injuries and illnesses related to inadequate workplace design requires a risk management approach combined with the application of engineering solutions [9]. Risk management begins with the identification and measurement of workplace ergonomics risk factors. After risk factors in the workplace are well understood, equipment and workstations can be redesigned to minimize exposure to the risk factors. The application of engineering solutions to reduce exposure to ergonomic risk factors requires a two-pronged approach: retrofit solutions are applied to existing work stations and equipment, and the design of new workstations and equipment encompasses ergonomic criteria (―fit once repeat many‖) [9].

24

2.2.1. Continuous improvements and breakthroughs

In order to optimize the processes, ergonomics should rely on continuous and breakthrough improvements. Continuous improvements are inexpensive, incremental changes, and made on a daily basis in order to drive out waste in operations. The power of continuous and breakthrough improvements through the application of ergonomics principles could be seen in the case of vehicle assembly plants of Toyota in Georgetown, Kentucky, United States. Namely, the plant faced with significant design changes between two model years, resulting in potential ergonomics risk and inefficiencies, particularly with the rear spoiler installation. The company hired a consulting firm to conduct an ergonomic risk assessment, work measurements analysis, and make recommendations for improvements [9]. The job process included nine specific detailed components, ranging from removing the trim panels, to attaching the spoiler to a final check and cleanup. The ergonomics team‘s risk assessment and operator interviews identified significant ergonomic risk. The operation, which required workers to climb inside the trunk, included such risks as extended reaches, deviated postures, and mechanical stress to numerous body parts. The team also performed a work measurements analysis of the entire operation. The cycle time of the operation totaled 20.4 minutes, with drilling, trunk preparation, and clean up accounting for more than half of that. The easiest of the recommendations was to decrease tool reach and design new hand tools for the wire harness installation. Tools were removed from the suspension system, placed in a mobile tool cart, and angled 15 degrees, reducing the awkward postures and extended reaches, resulting in reduced cycle times. A newly designed tool allowed operators to install the harness from outside the trunk, thus eliminating wasted motions, reducing ergonomic risk, and saving1.8 minutes [9]. The recommended breakthrough improvements included changing drill bits and designing an additional trunk part. By switching from a standard drill bit to a multi-flute end mill bit, operators were able to drill a hole through multiple layers of sheet metal and eliminate the need to vacuum up metals havings. This recommendation significantly reduced awkward postures and resulted in a time savings of 8.9 minutes. Due to the added weight of the spoiler, the trunk lid would not stay open, which resulted in the operator climbing into the trunk to replace the torsion rods with higher tension rods. Rather than replacing these rods, they designed a compression spring that was inserted into both hinges of the trunk, increasing the force required to closethe trunk lid. This eliminated non-value added motions and wasted material, significantly reduced ergonomic risk, and resulted in a savings of 2.0 minutes. The impact of these changes resulted in greater operation efficiency and time savings at the highest-risk operations. Toyota‘s $9,000 investment resulted in increased through put and 25

less waste. The task cycle-time was reduced from 20.4minutes to 7.7 minutes, a 62 percent reduction, translating into a projected annual savings of $262,000 in direct labor cost. Long-term savings include decreased medical and workers‘ compensation expenses, lower absenteeism, and reduced training of new associates [9].

2.3. Work related musculoskeletal disorders

Several work activities, in particular those associated with repetitive movements and with considerable level of stressor with extended assumption of uncomfortable posture,might be correlated to the insurgence of Muscolo-SkeletalDisorders (WMSDs= Work related MuscoloSkeletalDisorders) [10].The lowback, neck, shoulders, and upper limbs are the body parts most subject to risk (e.g., Roquelaure et al., 2006;Viikari-Juntura &Silverstein, 1999). Among the factors associated with the risk for developing work-relatedMSDs are individual, physical workplace, organizational,and psychosocial factors, like repetitive motion, forceful exertions, vibration and awkward postures. [11,12]. Figure 2.1 illustrates a systems model for conceptualizing the various elements of a work system that may affect musculoskeletal health. Working conditions can exert loads on workers which can lead to musculoskeletal strain. In this model, these various elements interact to determine the way in which work is carried out, and the effectiveness of the work in achieving individual and organizational needs and goals. At the center of this model is the individual with his/her physical characteristics, perceptions, personality and behaviors. The individual has technologies available to perform specific job tasks. The capabilities of the technologies affect performance and also the worker‘s skills and knowledge needed for its effective use. The task requirements also affect the skills and knowledge needed. Both the tasks and technologies affect the content of the job and the physical demands the job puts on the person. The tasks, with their technologies, are carried out in a work setting that comprises the physical and the social environment.

26

FIGURE 2.1: Conceptual model of the development of work-related musculoskeletal disorders There is also an organizational structure that defines the nature and level of individual involvement, interaction and control, and which determines workload and working hours. All of these work considerations combine to define the loads on the person. It is the person‘s capacity to deal with these loads that produces misfit that can lead to musculoskeletal strain. The ability to `balance‘ these loads determines the extent of misfit and potential strain [12]. It is known that these kinds of disorders involve high costs linked to absence, medical insurance, and rehabilitation, Figure 2.2. In addition to the overall increase of WMSDs, there is substantial evidence that ergonomics improvements result in financial gains for companies [11].

FIGURE 2.2: Cost analysis by occupational injury and illness types in Canada Therefore, the prevention of musculoskeletal disorders in the workplace often focuses on evaluation of jobs in terms of physical job requirements and the development of solutions to reduce the frequency, duration and / or intensity of the biomechanical stressors during work. As it was previously mentioned, ergonomics job analysis (EJA) is performed systematically: 27

-

Identify hazardous work situations that present peak exposures (e.g., extremely heavy lifting); Determine whether exposure is beyond conventionally recommended levels (e.g., working in knee ling postures for more than 2 hours per 8-hour shift); Rank order jobs for intervention; Evaluate whether interventions actually reduce exposures associated with jobs.

Common questions that arise in ergonomics practice include: -

Is greater than 50 lb ( 23 kg) ever handled during a job? Do workers in a job spend more than 25% of the work time in trunk flexion? Which tasks or jobs are the most hazardous? Did the lifting aid reduce the occurrence of manual materials handling or awkward body postures?

To obtain exposure information on various physical and environmental factors of a job, observational approaches—either in real-time or video—are often used. Such approaches allow many potential risk factors to be evaluated quickly; are generally less costly than instrumentation methods used in research; and may reduce the risk for information bias that appears to be present in self-reporting of ergonomics exposures. One popular observational job analysis approach involves the identification or evaluation of ergonomics risk factors from video recordings. The advantage of this approach over observational assessments made in real time is that videos can be replayed, or played in slowmotion or freeze-frame for a more careful assessment of exposures. This may reduce the risk of misclassifying exposures. On the other hand, video-based observations are generally more time consuming than other observational approaches [13]. Many ergonomic measures reduce the exposure to musculoskeletal disorder risks. It is difficult, however, to determine the effectiveness in reducing the actual MSD rates, because: There often is no discrete onset, -

The exposure and morbidity history accumulated before the intervention may still affect outcomes after the intervention, The recurrence of an earlier episodeafter the intervention may cloud the benefits, and

The number of workers involved is often too smallto provide statistical evidence [11,12]. A primary purpose of workplace interventions to control musculoskeletal disorders is to reduce the stress load to eliminate strain. As discussed below, this can be done by modifying the elements of the work system shown in Figure 2.16. Another tactic to control musculoskeletal disorders is to increase the capacity of the individual to handle greater loads, thereby reducing the possibility of a misfit. There are a variety of actions that have been applied in the workplace for eliminating or reducing the occurrence of occupational musculoskeletal disorders. These include engineering redesigns, changes in work methods, administrative controls, training, organized exercise, and using personal protective equipment to reduce exposures. In addition, work hardening and medical management are used to improve the 28

capacity of the person, or to treat symptoms. Some of these actions have been evaluated in research studies using both laboratory and field settings [12]. An example for the workplace interventions would be a replacing the large tables with two parallel workstations, where the majority of the required assembly steps are performed by two operators, and a third station in the series, where the products are finished and packed by one operator (Figure 2.3). [11].

FIGURE 2.3. Former batchwise production (left) and one-piece flow production (right) in final assembly. All three workstations, with the main elements are presented in Table 2.1 were ergonomically well designed. TABLE 2.1.Former and New Situations in Final Assembly and Print Assembly and the Main New Elements

29

In this case, it was considered that the decrease in physical load due to the variety of workstation adjustments to be significant, and on that basis it was estimated that up to30% of the work-related MSDs might be prevented. Prior to the intervention the sick leave percentage of 7% was observed, while after the intervention, it was decreased to 6.6 %. As a result of these changes, the company made a comparison between the costs made to replace workers because of sick leave before and after the intervention. It was observed that prior to the intervention, each worker under observation) had 1 day of overtime per month to compensate for sick leave for the entire population of workers, but after the intervention the need for overtime was totally eliminated. This equals €50,558 of annual savings due to reduction of number of sick leaves (based on gross salaries and a 30% overtime bonus). At the same time not just the cost-effective benefits are observed, but the increased productivity as well. Productivity increased significantly both in print assembly (by 20%) and final assembly (40%) [11].

2.3.1. Standing and walking posture during assembly operations

Standing is a common working position required by most occupations. Standing with in confined areas for extended periods of time intensifies musculoskeletal fatigue and body discomfort. During the last two decades, several health problems connected to prolonged standing have been reported in literature. These include lower extremity fatigue, pain, swelling and discomfort, venous blood pooling, low-back pain, and whole-body fatigue. High incidences of low-backpain have been associated with prolonged standing of 4 hours or more per day. Prolonged standing can also cause the joints in the spine, hips, knees, and feet to become temporarily immobilized or locked [14]. This immobility can subsequently lead to rheumatic diseases due to degenerative damage to the tendons and ligaments Machine operators, assembly-line workers, and others whose jobs are characterized by prolonged standing commonly report these problems. Ryan (1989) reported high incidences of lower extremity pain and discomfort among checkout personnel of supermarkets, who are confined to stand in restricted areas. The impact of such amplified fatigue, discomfort, and musculoskeletal disorders may lead to reduced productivity apart from other physiological and psychological stresses [14]. Various solutions to reduce these problems have been proposed in literature. Most studies have relied on ergonomic measures such as flooring and external aids such as mats, shoes, and sit/stand chairs. Standing posture is another influence able measure that has not received much attention. This study seeks to find if a dynamic standing posture produces less fatigue and more comfort than a stationary standing posture. Dynamic standing posture is an ergonomic posture in which a worker intermittently walks while he is on the job. Its stationary counterpart, stationary standing posture, is one in which the worker does not engage in walking breaks. A dynamic standing posture reduces the amount of time spent standing within a confined area, by engaging the legs in an intermittent motion within a wide 30

floor space. Several medical problems have been associated with long-term standing and an effective standing posture may offer a viable solution [14]. Fatigue related to standing has been quantified by objective measures, including electromyography muscle responses, performance measurements, skin temperature, and venous pressure. During prolonged standing tasks, postural movements are performed to diminish discomfort caused by psychological and physiological factors. Adjustments to body position are performed naturally to maintain comfort. These displacements are performed unconsciously and cannot be associated with any external source. Center of pressure (COP) is a common variable to study postural changes. COP displacements, which follow a fractal stochastic process, have been shown to correspond to the responses of the postural control system. Frequency of postural changes indicates fatigue since they are performed to avoid discomfort and fatigue. A positive relationship from neuromuscular fatigue to posturals ways has also been reported. Balasubramanian et al. [14] during their study used COP extracted features to evaluate fatigue. While standing, the weight-transmitting area of the foot (ball of great toe and heel) is compressed and deformed by the overlying pressure. This could inhibit the supply of blood, resulting in a deficiency of oxygen to tissue, and could cause discomfort or fatigue. For this reason, the quantum of foot pressure or the pressure distribution at the foot/floor interface may be used as an objective indicator of comfort or fatigue. To analyze foot pressure, two evaluation parameters of pressure were considered: (a) contact area as a function of pressure range, which is a time independent measure of pressure; and (b) area pressure change root mean square (aPcrms), which analyzes the dynamic behavior of pressure. aPcrms is also an indicator of comfort index (CI). Both these parameters were used to compare the CI between the two standing postures. A positive influence of the dynamic standing posture on standing comfort can be used to improve industrial productivity [14]. A significant and positive difference between dynamic and stationary standing during assembly work could be used to improve workplace ergonomics, industrial productivity, and occupational safety. These results may also influence the ergonomic design of shop floors, where in a particular job, such as assembling, may be distributed. The inferences can be generalized to all types of jobs where a prolonged standing posture is in practice [14].

2.3.2. Upper Extremity Assessment Tools

American Conference of Industrial Hygienists (ACGIH) is an association which is committed to prevent workers from occupational diseases. Hand activity levels (HALs) are introduced by the ACGIH for mono-task jobs performed longer than 4 hours per day. ―Task‖ which represents duty in ergonomics is different from ―task‖ representing the basic indivisible element in ALBP. Workers who repeatedly perform the same exertions every work cycle are

31

considered as acting ―mono-task‖ by definition [7]. HAL are offered for assessing the WMSD risks in such cases. The scale for HAL was proposed by Latko et al. [15]. and this scale range is from 0 to 10. In this particular scale, 0 represents ―completely idle‖ and 10 stands for the greatest level of ―continuous exertion‖. HAL is not only a function of frequency but also a function of work speed. There are two ways to determine HAL: 1. HAL can be obtained from Figure 2.4, which is based on the frequency of hand exertions, the idle time of hand exertions and the speed of motions and 2. HAL can be calculated from Table 2.2, which contains the exertion frequency and duty cycle.

FIGURE 2.4: Hand activity level (0-10) can be rated using the guideline

TABLE 2.2 Hand activity level (0-10) is rated based on exertion frequency and duty cycle (% of work cycle where a worker‘s force is greater than 5% of the maximum) [16]

32

Threshold Limit Value (TLV) is the guideline used by some industrial hygienists in trying to prevent occupational hazards [16]. By the applications of TLVs, most TLV methods fall into two categories. TLVs for physical agents are used in determining the exposure level of vibration, radiation and heat/cold stress, in which operators can work for certain time day after day nearly without suffering occupational diseases during and after career. Similar to TLVs for physical characteristics, TLVs for chemical substances help decide the safe levels of exposure to chemical substances. ―Threshold limit values for chemical substances and physical agents & biological exposure indices‖ are recommended [16]. HAL and normalized peak force (NPF) are the two dependent variables used in the ACGIH HAL TLV. Hand-arm (segmental) vibration is also a factor affecting WMSD risks, and TLV of vibration is introduced by ACGIH. [7].

2.3.3. Hand Vibration Exposure on an Assembly Line However, a lot of studies performed are motivated by management‘s concern about reports of cumulative trauma disorders in the upper extremities of workers who use pneumatic screwdrivers. The term ―vibration syndrome‖ is often used collectively for the symptoms associated with prolonged and repeated exposure to hand tool vibration. These symptoms include vasomotor irregularities, bone alterations and joint deformations, neurological disturbances and soft tissue damage such as changes in the chemical composition of the blood [17]. The condition known as ―Vibration Induced White Finger‖, ―Traumatic Vasospastic Disease‖ and ―Raynaud‘s Phenomenon of Occupational Origin‖ is a disease of the fingers and hands with both neurological and peripheral vascular symptoms. The Taylor and Pelmear classification system [18, 19] of the hand and arm condition is based mainly, on a patient‘s medical history of attacks of vibration white fingers. Epidemiological evidence has shown that there is a positive correlation between the total time of exposure and the severity of vibration white finger which suggests that the traumas are cumulative. [20] Bone deformity is another of the lesions caused by vibration exposure. Radiographic studies of the wrist elbow, shoulder and the cervical vertebrae of workers with weakened grip revealed abnormal findings in the elbows. Neurological disorders, such as disturbances in the sense of touch, increased sensitivity to cold, as well as ulnar paresis and paralysis, acroparesthesia accompanied by cramp-like pains in the extremities, and fatigue have been observed in workers exposed to vibration [17, 21]. Abnormal findings in nerve conduction velocity studies have been reported when symptoms of parasthesia, numbness and or pricking sensation were present in workers who had used vibrating tools for several years [22]. The international Standards Organization (ISO) exposure guidelines for hand-transmitted vibration [23] is a provisional consensus standard based on data from both practical 33

experience and laboratory experimentation derived primarily from subjective human response to hand-transmitted vibration and mechanical behavior of the hand-arm system. Exposure zones for hand-transmitted vibration are recommended for regular daily occupational exposure over long periods of time in terms of acceleration measured in one-third octave band spectra. Accurate estimation of daily vibration exposure for individual workers on the production line is difficult because of the highly variable operating techniques used between operators and anomalous contributors common to this operation such as cross threaded screws and back-up of co-workers who have fallen behind. Radwin and Armstrong [24] assessed vibration exposure by sampling directly the vibration produced from the specific tools of individual operators and prediction of daily exposure using in-plant data. They investigated the factory which produces small electromechanical appliances [24]. Twenty –three workers were employed in the assembly operation where seventeen of these positions required the use of pneumatic screwdrivers. One position necessitated operation of o two screwdrivers. The production rate was 2.9 units per minute which was below average with a production rate of 3.2 units per minute considered to be typical for the line. Five workers were new and inexperienced. Three additional workers were positioned on the line to back up the trainees. The screwdrivers used were in-line pneumatic screwdrivers measuring 21.6 cm long, 2.5 cm in diameter, weight 450 g. And are push-to-start, torque controlled tools designed to start when the bit or socket is depressed and stop when this pressure is removed. A clutch mechanism causes the rotor to slip when a screw is tightened to the set torque. The torque is in the 0.9 nm range and the tool was operated at 620 kPa air pressure [24].

FIGURE 2.5: A representative work station. Workers seated around a conveyor load their screwdrivers with screw scattered along a tray with a perforated bottom situated along side of the workbench. The loaded screwdriver is moved into position over a threaded hole with one hand while a wire with a crimped lug is positioned under the screw head with the other hand and screw installed.

34

Like most air powered screwdrivers, these are supplied with bare metal handles. Consequently many workers cover their handles with adhesive tape in an attempt to make the handles more comfortable. Most of the workers on the line wear light cotton gloves with the fingers removed. Many also cover their fingers with tape [24]. Figure 2.5 illustrates a representative work station. Hand and waist posture is not a problem since workers operate the in-line tools on a horizontal surface at elbow height while seated at a workbench, where the bench height is 97 cm from the floor. Two experiments were performed to assess the vibration exposure hazards. The first was an in-plant study to estimate the quantity of vibration exposure for each worker. The daily eight hour exposure was predicted from observed vibration samples for each position on the assembly line. The second experiment was a laboratory study to assess the quality of vibration exposure. TABLE 2.3: Mean and Eight Hour Vibration Exposure During Run and Clutch Cycles for Positions Requiring Operation of Power Screwdrivers

35

The average running and clutch slippage time (Table 2.3) were computed for each worker whose position required operation of a pneumatic screwdriver on the line. The number of observations ranged between 16 and 336 depending on the job of the individual operator. The screwdriver run time included engagement of the screwdriver during pick-up, screwdriver operations and operation of the screwdriver when screws were defective (requires reverse operation of the screwdriver to remove the defective screws). The time that the screwdrivers operate in the run and clutch phases, are dependent on several individual operator factors. Operators load their screwdrivers by pressing the tool over a screw which usually activates the push-to-start pneumatic motor [24]. The total eight hour exposure to vibration for workers on the line during the running portion of the screwdriver operation cycle ranged between 2.6 and 51.8 minutes (Table 2.3). The total eight hour exposure to vibration produced during clutch slippage ranged between 1.2 and 14.4 minutes. Six of the 17 positions (positions 4, 5, 6, 7, 8 and 9) (Table 2.3) approached or exceeded 30 minutes of total exposure to vibration produced during running. These jobs involved installation of the greatest number of screws on the line. The position 14B required the installation of a screw that was longer than the others, so the mean exposure time in the run cycle was higher (258 ms).

FIGURE 2.6: Center frequency of third octave bands

36

One-third octave band recordings for the x, y, and z axes in the run and clutch screwdriver and in the run phases for the automatic shut-off screwdrivers, shown in Figures 2.6, where within the ISO proposed two to four hour exposure boundaries. The greatest one-third octave band frequency component was at 31.5 Hz. The maximum predicted eight-hour exposure for the workers observed during the run phase was less than one hour. Thus, all the workers observed were subjected to vibration within the ISO guidelines for vibrations produced during the run phase of the screwdriver operation cycle [24]. Accurate estimation of worker vibration exposure is necessary when comparing alternative tool vibration signatures and using the ISO guidelines as a bench mark. Since hand tool operation is subjected to a high degree of variability in many types of assembly operations, the need to implement vibration exposure measurements by direct sampling is indicated to accurately estimate vibration exposure. Zhan Xu [7] and his collaborators, in their research applied a new approach in order to improve the conventional practice by which design an assembly line and reduce ergonomic risks. A study was conducted to design an assembly line for consumer electronics appliance (blender) assembly. Their research develops more explicit integration of task assignment and ergonomic measures using linearized formulations, which enables not only to integrate ergonomic task characteristics directly but also to use efficient solution methods to optimize assembly line design, in particular for upper body extremities. Moreover, these linearization methods can generalize ergonomic measures in table listed forms and incorporate them into other assembly line design problems [23]. The linear models developed in the research performed by Zhan Xu [7] help reduce WMSD risks. The ergonomic measures incorporated with task assignment can be used to limit the peak force and vibration exposure levels of upper body extremities in all stations in an assembly line. The developed methodology can be extended to incorporate a variety of ergonomic characteristics of assembly tasks. This feature provides flexibility to consider other ergonomic constraints during assembly line design. Hence, the research by Zhan Xu will help to prevent WMSD risks among assembly workers. So, their attempts [14] were: -

-

To develop a methodology that incorporates hand activity and hand-arm vibration measures into the task assignment models in assembly line design. In particular, to build linear functions integrating the exertion frequency, duty cycle, normalized peak force (NPF), vibration acceleration and vibration duration measures into the linear task assignment model. Next, these linear models were integrated for assembly line design by using mixed-integer programming (MIP). To demonstrate the feasibility of the modeling methodology to identify the impact of the ergonomic consideration on assembly line design in terms of task assignment, and examine the trade-offs between productivity and ergonomic conditions. The models developed in this research will help control WMSD risks with reducing possible negative impact on line efficiency.

37

The overall structure of the research was conducted as it is shown in Figure 2.7 First, the physical exposure in an assembly line for this research consists of five ergonomic characteristics: peak force, number of exertions, duty time, equivalent accelerations of vibration and vibration duration. A guideline for industrial hygienists concerning physical exposures of operators [16] was used for building the equations for the ergonomic characteristics. These numerical ergonomic representations mainly focus on upper body extremities. Second, these ergonomic formulas are created as linear functions of task assignment. These linear formulas allow the easier integration of the ergonomic measures into linear assembly line design models. Third, an assembly line design model is built by incorporating the linearized ergonomic measures and conventional assembly line characteristics into an MIP model. The conventional assembly line characteristics include task precedence, cycle time and the number of work stations. The objective function of the MIP model is to minimize the number of workers in the assembly line. Fourth, through numerical experiments, this research analyzed the effect of different ergonomic considerations by solving the MIP model with different combinations of ergonomic constraints. This analysis demonstrated the effectiveness of the new integrated approach compared to a conventional assembly line model without ergonomic considerations [7].

FIGURE 2.7: The structure of the research

38

3. THE RELATION OF TASK ASSIGNMENT TO EXERTION FREQUENCY, DUTY CYCLE, NORMALIZED PEAK FORCE, VIBRATION ACCELERATION AND VIBRATION DURATION This chapter describes how to incorporate task assignment and upper extremity ergonomic measures by equations.

3.1. Determination of hand activity and peak force measures by task assignment Task assignment is the procedure of dividing the workload needed for assembling one product into several elements and distributing them among work stations, and strongly affects individual worker‘s ergonomic condition. Every task possesses ergonomic characteristics affecting the hand‘s exposure levels such as the number of exertions, duty time and peak force. Because a worker‘s job in an assembly station is usually a set of several indivisible tasks, different task assignment (different sets of tasks assigned to each worker) results in different number of exertions and duty times for each worker. The relations of the exertion frequency and duty cycle to task assignment are formulated as linear Eqs. (1)-(2), respectively. In his research, Zhan Xu [7], assumed that one worker charges one work station‘s job, and does not share the job with other workers. Thus, a worker is equivalent to a station in terms of task assignment, and they are used inter changeably in this research. All tasks are performed in standing posture.

(1)

(2)

Equation (1) represents the exertion frequency at hand h of worker s, defined as the number of exertions per time unit. Eq. (1) expresses the sum of the numbers of exertions of all tasks assigned to workers divided by the cycle time of the station. In the equation, xis is the taskstation assignment variable, neih is the number of exertions of task i, hand h, CT is the cycle time in the assembly line. Thus, ef sh represents the exertion frequency in station s, hand h. Eq. (2) represents the duty cycle at worker s hand h, defined as the ratio of total time of duties in a work cycle. Eq. (2) expresses the total duty time of all tasks assigned to workers divided by the cycle time of the station. In the equation, xis is again the task station assignment 39

variable, dtih is the duty time of task i hand h, CT is the cycle time in the assembly line. Thus, dcsh represents duty cycle in station s, hand h. In these equations, number of exertion ( neih )

represents the number of energy costing motions, and duty time ( dtih ) is the duration where the worker exerts more than 5% of the maximal force [16]. Both of them are measured based on single assembly task. Peak force is also an important factor contributing to WMSD risks [25-27]. A worker‘s maximal force performed in one repeated job is represented as peak force and usually normalized as a dimensionless measure to fit the general population [28, 29]. If tasks with large amount of exertions and heavy normalized peak force (NPF) are assigned to a worker, this worker‘s WMSD risk could be considerably high. The normalized peak force limit is related to the combined level of the exertion frequency and duty cycle (expressed as hand activity level at (Table 2.2)), and this relationship is explained below. In Figure 3.1, the horizontal axis represents the hand activity level (HAL) and the vertical axis the NPF. According to a variety of workers‘ strengths, an action limit (AL) is recommended as a safe bound. One expression of the AL was proposed as shown in Eq. (3) [30] and is shown as the dashed line in Figure 3.1. The combination of HAL and NPF levels should be below the AL line. Thus, NPF is restricted based on the HAL as shown in Eq. (4).

FIGURE 3.1 The TLVs based on HAL and NPF.

40

(3)

(4)

where npfl sh = NPF limitation of station s and hand h, halsh = HAL value of worker s and hand h, and npf sh = NPF value of worker s and hand h.

3.2. Determination of hand-arm vibration measures by task assignment As we can see from the previous chapter vibration duration and vibration equivalent acceleration are also related to upper extremity WMSDs. The daily vibration duration and its relation to task assignment are given in Equation (5).

(5)

Eq. (5) represents the sum of the vibration duration of all tasks. In the equation, xis is the taskstation assignment variable, vti = the duration of vibration of task i, NC is the number of work cycles per day in the assembly line, defined as the total daily working time divided by the cycle time. Thus, s TD represents the daily vibration duration in station s. The frequency-weighted, equivalent, component acceleration of a set of tasks is expressed by Eq. (6), and the largest equivalent acceleration among three directions is considered as the dominant acceleration used to evaluate the vibration level [16].

(6)

o where akeq , s = the equivalent, frequency-weighted component acceleration of station s,

direction o, akio = the frequency-weighted, rms acceleration of task i, direction o, TDs = the total daily exposure duration of vibration of station s, and NC = the number of work cycles a worker completes per day [16]. Just for now, note that Eq. (6) is not a linear function of task 41

assignment. Because the vibration data of each task are measured in the realistic task operation, the data usually include the impacts of grip forces on vibration. Also, the effect of static grip force on vibration is known not significant [31]. Thus, the TLVs of HAL and vibration acceleration can be evaluated independently [7]. The vibration equivalent acceleration TLVs are shown in Table 3.1. TLV for acceleration is the limit of the largest acceleration in all three directions. Any equivalent vibration acceleration in the station should not exceed the TLV. TLVs vary by the daily vibration durations in the station as shown in Table 3.1.

TABLE 3.1: TLVs for exposure of the hand to vibration in either X, Y, Z Axis [16]

* The total time vibration enters the hand per day, whether continuously or intermittently. ** Typically, the accelerations of one axis show dominion to those of the other two axes. *** g = 9.81 m/s2

4. OPTIMIZATION MODEL

This section describes the optimization model for assembly line design. This model is to minimize the number of workstations while considering production rates, hand activity and hand-arm vibration. The main decision variable represents the task assignment to stations/workers. The models include two types of constraints: ergonomics and conventional assembly line design constraints. The ergonomic constraints include hand exertion frequency, duty cycle, normalized peak force (NPF), vibration acceleration and vibration duration. The conventional assembly line design constraints include cycle time and task precedence constraints. The optimization model is presented as follows.

42

Min z

(15)

Subject to

(1)

(2)

(7)

(8)

(9)

(10)

(11)

(12)

43

(3)

(4)

(5)

(14)

(17)

(18)

(19)

(20)

(21)

The explanations of the models are as follows. The objective function, Eq. (15), is to minimize the number of workers (workstations). Constraint (1) indicates the exertion frequency is derived from dividing the total number of exertions by the cycle time. Constraint (2) shows the duty cycle is linearly dependent on the duty time of each task. Constraints (7) can determine Table 4.1‘s row index based on exertion frequencies, and duty cycles can determine Table 4.1‘s column index by Constraints (8). Constraints (7)–(8) are explained as follows: The equation (7) gives the row number ( ) in Table 4.1. The equation (8) relates the duty cycle ( ) attained from task assignment and the corresponding representative duty cycle used in the look-up table (Eq. (8) and Table 4.1). This equation gives the column number ( ) in Table 4.1. Refer to Table 4.2 and Table 4.3 for the parameters in the equations. [7] 44

The row and column numbers obtained by Eqs. (7)-(8) are used to select the corresponding HAL value. The linear equations (9)-(10) determine value of the HAL at each station using the row and column numbers from Eqs. (7) and (8) and the HAL value table (Table 4.1). A binary variable is defined to indicate the cell position in the HAL table. If the cell u is selected for station s then Then the equations (9) – (12) determine the value of . Eq. (9) determines the address in the HAL table (Table 4.1). Variable alone can represent the addresses of HALs. Equation (10) ensures only a single cell is selected from Table 4.1. Eq. (11) finds the only cell in Table 4.1 that should be used (u for which should be one). Eq. (12) assigns the corresponding HAL values using the value determined by Eq. (11). HTu represent the value of HAL when in Table 4.4. The linearization of the conventional look-up table based procedure is possible by converting the two-dimensional relations shown in Table 4.1 and development of new equations. [7]

TABLE 4.1: Reference table of HAL with m = 5 [16]

ACGIH TLV is applied to set the limits of ergonomic measures in this research. The table considers the worst case for each unreachable entry of Table 2.2, so the results may be conservative. Such cases, however, will be rare.

45

TABLE 4.2: The set of ranges of exertion frequencies [16]

TABLE 4.3: The set of ranges of duty cycles [16]

TABLE 4.4: The conversion of intermediate value

to HAL

Constraint (3) represents AL. Constraint (4) shows NPF of a worker should be always less than AL. Constraint (5) represents the total daily vibration exposure duration of worker s. The acceleration limits are also shown in linear formulas. Eq. (6‘) is a linear form of Eq.(6) [16], defining the equivalent frequency-weighted, rms, component acceleration. Eq. (13) shows the equivalent accelerations of workers Should be smaller than the acceleration TLVs ( ). Equation (14) is derived from Eq. (6‘). Thus, constraint (14) ensures handarm vibration accelerations.

46

Constraints (17)-(20) are the traditional constraints in assembly line design. So called the ―ghost task‖ is defined so that the task requires all the other tasks should be done before it. Thus, the ghost is always in the last station. All characteristics in the ghost task are zeros, so it will not affect the solutions of models. Constraint (17) ensures that the last station contains the ghost task for the precedence relation and the last station number is the total number of assembly line stations (workers). Constraint (18) shows every task can only be assigned to only one station once. Constraint (19) makes the sum of task times for each station under the cycle time. Constraint (20) makes sure the sequence of work stations does not conflict with task precedence. This optimization function has the following characteristics. First, it is a mixed-integer linear program. Thus, (1) the linear form of the constraints allows faster calculation than other non-linear constraints found in the literature, (2) the ergonomic data tables used for the constraints can be conveniently replaced with more sophisticated tables or other ergonomic measure tables, and (3) other linear assembly line constraints can be incorporated in this optimization program easily. Second, the computational complexity of the optimization formulation is not significantly higher compared to general assembly line design formulation. [7]

TABLE 4.5: The conversion of daily vibration to acceleration TLV [16]

47

5. NUMERICAL EXPERIMENTS AND DISCUSSION

To verify if the models effectively control hand activity and vibration received by the upperextremities in assembly line design, a numerical experiment were conducted by the researchers. The solutions from traditional model and ergonomics considered model are compared [7].

5.1 Manufacturing Task Description and Parameter Estimation

Task

Description

1. Placing Control Box on Body Roof by Automatic Shut-off Screwdriver 2. Placing Motor Base on Body Roof by Automatic Shut-off Screwdriver 3. Putting Magnet on Motor Base by bare Hands 4. Placing Motor Top on Motor Base by Automatic Shut-off Screwdriver 5. Fixing Motor Top on Motor Base by Clutch Screwdriver 6. Connecting Wires to Control Box by Plier 7. Fixing Moto Base to Body Roof by Clutch Screwdriver 8. Connecting Wires to Moto by Plier 9. Connecting Fan with Motor by Wrench 10. Connecting Body Bottom and Body Feet by automatic Shut-off Screwdriver 11. Fixing Control Box on Body Roof by Clutch Screwdriver 12. Pasting Label by Hands 13. Fixing Body Bottom by Clutch Screwdriver 14. Fixing Top Rod by Hands The numbers in the circles represent assembly tasks. 1 and 2 represent vibration Profiles 1 and 2, respectively. FIGURE 5.1 The task description and precedence of assembly tasks with vibration profiles

48

A case study was conduct to design an assembly line for consumer electronics appliance (blender) assembly. The assembly process consists of 14 assembly tasks (Figure 5.1). The hand activity data for these tasks are estimated from the assembly task analysis in IMSE 898 and vibration data are estimated based on a study by Radwin and Armstrong, and Xu and Hao [24, 32]. All tasks are assumed to be performed in standing posture. Data related to HAL and vibration are shown in Table 5.1. Normalized peak force (NPF) is obtained by the methodology introduced in part which is passed above based on Table 5.2. The peak force applied in this case study is limited to grip force. These data were created by analyzing assembly tasks. The data is used as a sample set of data to initiate the model testing. The measured times, and an estimate of the performance times as input data for the numerical experiments, are used. The data, in this research, based on one subject, and multiple subject tests were be used in a future study to generate recommendations for practical use. [7]

TABLE 5.1: Data of exertion number, cycle time and NPF Task No. 1

2

3

4

5

6

7

8

9

10

11

12

Task Time (Second)

9.3

23.8 3.6 12.7 11.1 30.9 17.1 16.7 18.9 11.5 14.3 10.3

Number of Exertions (R*)

3

6

2

3

3

Number of 6Exertions (L)

1

2

2

1

3

3

5

Duty Times (R)(Second)

5.8

7

2.2

8.3

6.7

3.3

11.4

Duty Times (L) (Second)

1.8

1

2.2

8.3

7.0

4.8 10.0 7.1

3

NPF of Task (R)

4.3

4.3

2.6

4.3

4.3

NPF of Task (L)

2.8

2.8

2.8

2.8

2.8

5

4

0

3

2

1

1

-

1

Vibration Duration (Second) Acceleration Profile

2

5

4.3

4

6

4

4.3

2

1 9.1

2 4.0

3.5

2.8

2.8

0

4

0

0

2

-

-

-

3

8.8

4.3

1.9

3

2.8 2.8

7 1

6

3

10.3 1.8

7.0

14

15.4 18.6

3

3

7.4

0.6

2.6

3

13

5.5 5.7

1.8

5.0 8.0

4.3 1.7

4.3

1.9 1.9

4.3

2.8 2.8

2

0

3

0

2

-

2

-

R “ and “L” represent right and left, respectively.

49

TABLE 5.2 Male power grip strengths in Newton (N)

Dominant (right) 463.5*

Non-dominant (left ) 398.9

Subject age 18-65

Population Office workers

532.1

474.3

18-65

Laborers

556.6

514.5

18-65

Skilled

589.0

532.1

18-65

Semi-skilled

Data with * mark represent the data used by Z. Xu during the research [7].

FIGURE 5.2 The accelerations and the coordinate figures of the two screwdrivers. 50

The acceleration profiles were determined as follows. Because pneumatic tools cause considerable vibration in assembly line processes, this case study assumed the use of pneumatic tools. It is assumed that two types of pneumatic tools, Clutch Screwdriver and Automatic Shut-off Screwdriver, are used during assembling processes. Profile 1 represents a task placing screw using Automatic Shut-off Screwdriver. This task contains lower acceleration vibration but longer vibration duration compared to those of Profile 2. Profile 2 represents a task using Clutch Screwdriver tightening screws. The profiles of tasks are shown in Figure 5.1. This task causes relatively severe vibration acceleration and short vibration duration. The accelerations and the coordination of the two screwdrivers are shown Figure 5.2. The diameters of the two screwdrivers are assumed the same. All accelerations are frequency weighted based on ISO 5349 and are explained in Appendix A. [7]

5.2 Results from Different Combinations of Ergonomic Constraints Four line design models (LBMC, LBMCH, LBMCV and LBMCHV) were compared. These models were different in terms of the different combination of hand activity and hand-arm vibration constraints included in the assembly line design formulations. The acronyms for the models are shown in Table 5.3. These assembly line design models were solved by a commercial MIP software package CPLEX® (IBM) by Zhan Xu. [7]

TABLE 5.3: Acronyms for the line balancing models.

Model

Combinations of Constraints

LBMC

Conventional constraints only

LBMCH

Conventional and hand activity constraints

LBMCV

Conventional and vibration constraints

LBMCHV

Conventional , hand activity and vibration constraints

The result summarized is as a schematic diagram in Figure 5.3, and the detailed task station assignment is included in Table B.1in Appendix B. The exertion frequency, duty cycle, NPF and its TLV are shown in Appendix B. The vibration acceleration and duration is shown in Table B.5 in Appendix B. As shown in Figure 5.3, LBMCHV (the model with all hand activity and vibration constraints) is the only case in which all ergonomic exposures are below the TLVs. This is due to the added constraints in LBMCHV. 51

FIGURE 5.3: Results from different constraint combinations

Figure 5.4 represents the HAL & NPF values and TLVs. Table 5.4 shows the relative performance of each task. Table 5.5 shows dominant accelerations and TLV. The detailed analysis shows the violation of constraints in each case. For example, Figure 5.4 (a) and (c) show at least one of HAL values exceeded the AL line in the models not considering hand activity (LBMC, LBMCV). Figure 5.4 (b) shows that all HAL values satisfy the TLVs.

52

HAL TLV of LBMC

HAL TLV of LBMCH

HAL TLV of LBMCV

HAL TLV of LBMCHV Figure 5.4 HAL TLV of LBMC, LBMCH, LBMCV AND LBMCHV. In each case, there are six points (each hand at three stations), but some points are indistinguishable because they have the same values in HAL and NPF 53

Table 5.4 shows considerable hand activity level change in LBMCH and LBMCHV, compared to LBMC: the stations with high HAL exposure in LBMC decreased HAL, whereas stations with relatively low exposure in LBMC increased HAL. HAL is more balanced in LBMCH and LBMCHV. Figure 5.4 also shows the vibration accelerations did not satisfy TLV in the models without considering vibration (LBMC, LBMCH). Table 5.4 shows LBMCH did not improve the vibration condition. Station 3, which already exceeded vibration TLV in LBMC, even suffered higher vibration exposure than that in LBMC. However, the vibration accelerations were well controlled in LBMCV and LBMCHV. These comparisons demonstrate that the lack of ergonomic considerations in line design may result in severe work conditions in terms of hand activity and vibration. This ergonomic problems in stations needs adjustment in working conditions in individual tasks or numerous trial-and-error based task switching between stations until all the ergonomic measures are satisfied. The new approach can help to overcome this problem.

TABLE 5.4: Relative value of ergonomic measures compared to LBMC case

The relative values are calculated as,

here variable represents

ef, de, npf, ak or dt, and model represents LBMCH, LBMCV or LBMCHV.

54

TABLE 5.5: Vibration Results The acceleration exceeds TLV. 1 2 3

10.8 8.5 12.6*

0.88 0.72 0.80

12 12 12

1 2 3

7.7 2.5 13.3*

0.88 0.24 1.28

12 12 8

1 2 3

10.8 9.6 11.5

0.88 0.56 0.96

12 12 12

1 2 3

10.8 11.9 9.8

0.88 0.72 0.80

12 12 12

55

CONCLUSIONS Ergonomic deficiencies in work place design hamper work performance and create high, frequently unilateral stresses which accelerate worker fatigue and can cause adverse effects on physical health over the longer term. They prolong task completion times and also lead to an increase in number and severity of worker errors. The result is loss of productivity, not only at individual worker level but also in the company as a whole.

Companies that embrace ergonomics expect a safe work environment as well as an increase in efficiency and productivity. It is the goal of ergonomics and human factors engineering to achieve an optimal fit between work requirements and operator capabilities. Unfortunately, new ergonomically designed tools are not always necessarily improvements based on direct physical measurement, user‘s perceptions as well as productivity. In the example we elaborated, the researchers proposed a combined productivity and ergonomics methodology for assembly line design to help reduce the WMSDs in upper-body extremities among assembly workers, and successfully developed a model to verify the feasibility of the methodology. Linear formulas integrating ergonomic measures (exertion frequency, duty cycle, normalized peakforce (NPF), vibration acceleration and vibration duration) into task assignment models were established. Through numerical experiments, the researchers analyzed the effect of selected different ergonomic considerations by solving the MIP (mixed-integer programming) models with different combinations of ergonomic constraints. This analysis of the numerical experiments demonstrated the effectiveness of the new integrated approach compared to a conventional assembly line model without the ergonomic consideration. Thus, the numerical experiments demonstrated the feasibility of the modeling. This implies that the researchers successfully incorporated several ergonomic measurements into assembly design models, and the models may control the exposure of hands without sacrificing line efficiency. By the integration, the gap between ergonomics and assembly line design studies was decreased. The linearized ergonomic constraint models allow more explicit integration of task assignment and ergonomic measures in line design. Thus, this approach helps to overcome the problem of the conventional assembly line design: - separated assembly task assignment, and - ergonomic evaluations. In addition, the linearized ergonomic measures enable the use of efficient solution methods for linear assembly line design models. This methodology also opens the possibility of incorporating other diverse ergonomic characteristics of assembly tasks. The ergonomic measures related to hand exposures in all assembly workstations are controlled by the developed constraints. This approach can be used to help in preventing WMSDs among assembly workers by optimizing task assignment. 56

The main aim of our literature survey was to show the cases of the models integrating ergonomics and productivity measures. However, the direct application of the elaborated linear model in industrial settings is limited by the assumptions made during the model development. The application of the developed model requires an adaptation to the real industrial situations, which could be the base for some future researches.

57

APPENDIX A THE PROCEDURES FOR CALCULATING FREQUENCY WEIGHTED ACCELERATION The frequency weighted acceleration is calculated by

and Figure A.2

and Figure A.1, where f K = the f-th one-third-octave band (weighting factors in Table A.1), f a = the measured acceleration in the f-th one-third-octave band (Figure A.2 and Figure A.1), and F = the number of one-third-octave bands (International Organization for Standardization.

FIGURE A.1 One-third octave band spectra for 3,4 cm diameter automatic shut-off screwdriver

FIGURE A.2 One-third octave band spectra for the 2,5 cm diameter clutch screwdriver in the slippage of the clutch condition 58

TABLE A.1: Frequency- weighting factors for hand-arm vibration

Central frequency (HZ)

Weighting factor )

Central frequency (HZ)

Weighting factor )

6.3

1.0

100

0.16

8

1.0

125

0.125

10

1.0

160

0.1

12.5

1.0

200

0.08

16

1.0

250

0.063

20

0.8

315

0.05

25

0.63

400

0.04

31.5

0.5

500

0.03

40

0.4

630

0.25

50

0.3

800

0.02

63

0.25

1000

0.016

80

0.2

1250

0.0125

59

APPENDIX B RESULTS OF LBMC, LBMCH, LBMCV AND LBMCHV Table B.1 Hand activity result of LBMC Station Tasks Assigned F1,F9,F11,F12,F13

Station 1 F2,F3,F4,F7,F8,F14

Station 2

Station 3

F5,F6,F10

Exertion Frequency (R)

0.26

0.08

0.18

Exertion Frequency (L)

0.20

0.08

0.18

Duty Cycle (R)

0.437

0.174

0.274

Duty Cycle (L)

0.368

0.188

0.18

HAL (R)

5

1

2

HAL (L)

2

1

1

NPF (R)

4.3

4.3

4.3

NPF (L)

2.8

2.8

2.8

NPF TLV (R) 4.4444

2.7778

5

NPF TLV (L)

4.4444

5

5

60

Table B.2 Hand activity result of LBMCH

Station Tasks Assigned

Station 1 F1, F2,F6,F11,F14

Station 2 F1,F4,F8

Station 3 F5,F7,F9,F10,F12,F13

Exertion Frequency (R)

0.21

0.09

0.22

Exertion Frequency (L)

0.15

0.07

0.17

Duty Cycle (R)

0.321

0.196

0.368

Duty Cycle (L)

0.244

0.176

0.314

HAL (R)

2

1

2

HAL (L)

2

1

2

NPF (R)

4.3

4.3

4.3

NPF (L)

2.8

2.8

2.8

NPF TLV (R)

4.4444

5

4.4444

NPF TLV (L)

4.4444

5

4.4444

61

Table B.3 Hand activity result of LBMCV

Station Tasks Assigned

Station 1 F2,F3,F4,F7,F8,F14

Station 2

Station 3

F1,F5,F6

F9,F10,F12,F13

Exertion Frequency (R)

0.26

0.09

0.17

Exertion Frequency (L)

0.20

0.07

0.12

Duty Cycle (R)

0.437

0.158

0.29

Duty Cycle (L)

0.366

0.136

0.232

HAL (R)

5

1

2

HAL (L)

2

1

1

NPF (R)

4.3

4.3

4.3

NPF (L)

2.8

2.8

2.8

NPF TLV (R)

2.7778

5

4.4444

NPF TLV (L)

4.4444

5

5

62

Table B.4 Hand activity result of LBMCHV

Station Tasks Assigned

Station 1 F2,F3,F4,F5,F6,F7

Station 2 F1,F5,F8,F11,F14

Station 3 F9,F10,F12,F13

Exertion Frequency (R)

0.19

0..19

0.14

Exertion Frequency (L)

0.13

0.17

0.09

Duty Cycle (R)

0.323

0.376

0.187

Duty Cycle (L)

0.263

0.327

0.144 2

HAL (R)

2

2

HAL (L)

2

2

1

NPF (R)

4.3

4.3

4.3

NPF (L)

2.8

2.8

2.8

TLV of NPF (R)

4.4444

4.4444

4.4444

TLV of NPF (L)

4.4444

4.4444

5

63

Table B.5 Detail vibration results

Daily Vibration Exposure Duration

Model

Acceleration Limit (m/ )

LBMC

1 2 3

10.8 * 8.5 12.6

2.4 2.1 2.6

3.1 2.9 3.4

0.88 0.72 0.80

12 12 12

LBMCH

1 2 3

7.8 1.8 13.3

2.0 1.5 2.7

2.9 2.5 3.5

0.88 0.24 1.28

12 12 8

1 2 3

10.8 9.6 11.5

2.3 2.2 2.5

3.2 3.0 3.3

0.88 0.56 0.96

12 12 12

1 2 3

10.8 11.9 9.8

2.4 2.5 2.3

3.2 3.3 3.1

0.88 0.72 0.80

12 12 12

LBMCV

Shaded cells represent the values in them are dominant accelerations

64

REFERENCES [1] N. Boysen, M. Fliedner, A. Scholl, A classification of assembly line balancing problems, Working and Discussion Paper Series School of Economics and Business Administration Friedrich-Schiller-University Jena, Publisher Wirtschaftswissenschaftliche Fakultat Friedrich-Schiller-Universitat Jena Carl-ZeiЯ-Str. 3, D-07743 Jena, ISSN 16111311, 12, 2006, 1-18, www.wiwi.uni-jena.de [2] L. Capacho Betancourt, ASALBP: the Alternative Subgraphs Assembly Line Balancing Problem. Formalization and Resolution Procedures, Doctoral Thesis, Technical University of Catalonia, 2007 [3] Ch. Becker, A. Scholl, Balancing assembly lines with variable parallel workplaces: Problem definition, model and, exact solution procedure, Working and Discussion Paper Series School of Economics and Business Administration, Friedrich-Schiller-University Jena, Publisher Wirtschaftswissenschaftliche Fakultat Friedrich-Schiller-Universitat Jena CarlZeiЯ-Str. 3, D-07743 Jena, ISSN 1864-3108, 6, 2008, 1-30, www.jbe.uni-jena.de [4] W.Grzechca, Final Results of Assembly Line Balancing Problem, Assembly line, Theory and Practice, 2011, Publisher: InTech, ISBN: 978-953-307-995-0, Available from: http//www.intechopen.com/articles/show/title/final-results-of-assembly-line-balancingproblem. [5] P. R. McMullen, G. V. Frazier, A heuristic for solving mixed-model line balancing problems with stochastic task durations and parallel stations, Int. J. Production Economics, 51, 1997, 177-190. [6] H-H Kao, D-H Yeh, Y-H Wang, J-Ch Hung, An optimal algorithm for type-I assembly line balancing problem with resource constraint, African Journal of Business Management, 4, 2010, 2051-2058. [7] Z. Xu, Design of assembly line with the concurrent consideration of productivity and upper exremity musculoskeletal disorders using linear models, master thesis, Industrial and management systems engineering, University of Nebraska- Lincoln, 2010. [8] Gabriele Winter, Karlheint Schaub and Kurt Landau, Stress screening procedure for the automotive industry, Occupational Ergonomics 6 , 2006, 107–120. [9] Scott Smith, Elements of effective ergonomics, Three basic steps to securing and sustaining ergonomics success, Industrial Engineering, 35, 2003, 49-52. [10] D.Battini, M. Faccio, A. Persona, F. Sgarbossa, Linking ergonomics evaluation andassembly system design problem ina new integrated procedure, 19th International Conference on Production Research, Topic: Human Factors, http://www.icpr19.cl/, 2007. [11] Michiel P. de Looze, et al. Cost-Effectiveness of Ergonomic Interventionsin Production, Human Factors and Ergonomics in Manufacturing & Service Industries 20, 2010, 316–323. 65

[12] Ben-Tzion Karsh, et. al, The efficacy of workplace ergonomic interventions to control musculoskeletal disorders, Theor. Issues InErgon. Sci., 2, 2011, 23-96. [13] Victor L Paquet, Svend Erik Mathiassen, P. G. Dempsey, Video-based ergonomic job analysis, A practitioner‘s guide, Professional Safety, 11, 2006 27-34. [14] V. Balasubramanian,K. Adalarasu, R. Regulapati, Comparing stationary standing with an intermittent walking posture during assembly operations, Human Factors and Ergonomics in Manufacturing, 18, 2008, 666–677. [15] W. A. Latko, T. J. Armstrong, J. A. Foulke, G. D. Herrin, R. A. Rabourn, and S. S. Ulin, Development and evaluation of an observational method for assesing repetition ın hand tasks, American Industrial Hygiene Association Journal, 58, 1997, 278-285. [16] ACGIH. 2008. 2008 TLVs and BEIs (ACGIH, TLVs and BEIs). American Conference of Govermental Industrıa. [17] J. Hasan, Biomadical aspects of low-frequency vibration, A selective review, Work, Environ., Health, 6, 1970, 19-45. [18] W. Taylor, Vibration white finger in the workplace, J. Soc. Occup. Med., 32, 1982, 159166 [19] W. Taylor, P. L. Pelmaer, Vibration white finger in industry, Acedemic Press, London, 1975, 17-22. [20] W. Taylor, P. L. Pelmaer, J. C. G. Pearson, Vıbratıon-induced white finger epidemiology, Vibration White Finger in Industry, Eds, W. Taylor, P.L. Pelmear, Academic Press, London, 1975, 1-13. [21] T. J. Partenen, T. Kumlin, M. J. Karvonen, Subjevtive symptoms connected with exposure of the upper limbs to vibration, Work, Envıron., Health, 7, 1970, 80-81 [22] A. M. Seppalainen, Nerve conduction in the vibration syndrome, Health, 7, 1970, 82-84.

Work, Envıron.,

[23] Internatıonal Organization for Standardization (ISO), Technical comittee 108Mechanical vibration and shock, Sub-comittee 4: Human exposure to mechnanical vibration and shock, Draft proposal for a second drfat ınternational Standard ISO/DIS 5349, Guide fort he measurement and assesment of human exposure to vibration transmitted to the hand, Secrateriate, F. R. Germany, 1980. [24] R.G. Radwin and T. J. Armstrong, Assesment of hand vibration exposure on an assembly line, Am. Ind. Hyg. Assoc. J., 46, 1985, 211-219. [25] T. J. Armstrong, Preventing waste due to fatigue and muscoskeletal disorders, MODAPTS News, 25, 2008, 3-4

66

[26] ACGIH. ACGIH offers publications and products for occupational health Professionalproduced detail a. Internet on-line, Available from: http://www.acgih.org/store/ProductDetail.cfm?id=652, [7/13/2010, 2010]. [27] ACGIH. TLV chemical subtances introduction on the uses of TLVs and BEIs b. Internet on-line, Available from: http://www.acgih.org/products/tlvintro.htm, [7/13/2010, 2010]. [28] ACGIH.2009. Hand activity level: TLV® physical agents 7th edition documentation. [29] S. Hignett, L. McAtamney, rapid entire body assesment (REBA). Applied Ergonomics, 31, 2000, 201-206. [30] P. Drinkaus, R. Sesek, D. S. Bloswick, C. Mann, T. Bernard, Job level risk assesment using task level ACGIH hand activity level TLV scores: a pilot study, International Journal of Occupational Safety and the Ergonomics, 11, 2005, 263-281 [31] I. Pyykko, M. Fakkila, J. Toivanen, O. Korhonen, T. Hyvarinen, Transmission and vibration in the hand-arm system with special reference to changes in compression force and acceleration, Scandinavian journal of Work, Enviroment and Health, 2, 1976, 87-95 [32] Z. Xu, G. Hao, IMSE 898 Term Project: Assembly of Oster 6640 Blender, 2009

67