Creating Visual Work Instructions to Ensure Safe and ...

Open Eng. 2016; 6:691–699

Research Article

Open Access

Matúš Beluško, Matúš Hegedüš, and Gabriel Fedorko*

Creating Visual Work Instructions to Ensure Safe and Fluent Operation of the Semi-automatic Production Lines DOI 10.1515/eng-2016-0097 Received Jun 23, 2016; accepted Jul 30, 2016

Abstract: The process of robotics in the world, but also in Slovakia is constantly increasing, and although new types of robots are being developed, the human hand of those work stations cannot be removed completely. Despite the already developed and functioning collaborative robots that allow safe operations side by side with humans, it is necessary for a worker to known accurate, fast and safe procedure of his working activities. Visual work instructions (VWI) are made for this purpose. The worker must be able to quickly understand the principle of work procedure and get quick information about potential risks in very short time. Therefore, great regard for the content of VWI must be taken. At first glance, a simple visual work instructions enable not only to increase safety at the workplace, but also increase workers productivity. Keywords: Standard work instruction; picture-based work instruction; developing flowchart for standard work instruction; types of work instructions; decision making in work instructison creation

1 Introduction Automation in production lines especially in automotive industry is already common thing these days. Such production lines consist of few to dozens of robots or fully automated machines [1, 2]. However, even in fully automated

*Corresponding Author: Gabriel Fedorko: Technical University of Košice, Faculty of Mining, Ecology, Process Control and Geotechnology, Letna 9, 042 00 Kosice, Slovakia; Email: [email protected] Matúš Beluško: Technical University of Košice, Faculty of Mining, Ecology, Process Control and Geotechnology, Letna 9, 042 00 Kosice, Slovakia; Email: [email protected] Matúš Hegedüš: Technical University of Košice, Faculty of Mining, Ecology, Process Control and Geotechnology, Letna 9, 042 00 Kosice, Slovakia; Email: [email protected]

production lines there is a need for human workers who provide input of material and output of finished products. Therefore these production lines or cells might be called semi-automatic production lines. The human effectiveness and reliability in such production lines is crucial for fluent material flow in a whole production system. To ensure high overall performance it is recommended to create work instructions for standard operation procedures. When creating a Work Instruction for job aid it is important to mirror the specific standard operating procedure or process. This means you need to fully understand the specific steps of the procedure, its phases or the tasks and its duties [3]. There are several approaches to achieve high human effectiveness in semiautomated productions. They differ from field to field but most of them agree that in one company should be only one standard template for all working instructions such as manufacturing, material handling, welding and others, in order to achieve not only high effectiveness of them but also to speed up their development [4–7]. In this work we focus on robotic cellular productions in automotive industry. At the end of research, our very own approach of creating work instructions template is made. According to [7], Standardized Work Instructions are instructions designed to ensure that processes are consistent, timely and repeatable. Often the SWI are printed and posted near the operator’s work station. The idea is that team leaders and managers should follow up if the operator uses and can use the instruction. The goals and actual results of using SWI are improvements in: • • • •

Quality of the finished product, Consistency of the finished product, Throughput of the process, Safety of the operator [7].

How to use work instruction: • Determine the activity that needs to be documented. • Determine the education and skill level of the target audience that will be accessing the instructions.

© 2016 M. Beluško et al., published by De Gruyter Open. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License.

Unauthenticated Download Date | 7/1/17 12:41 PM

692 | M. Beluško et al. • Develop a format for the work instructions. This may include the name of the author, the name of the document, the date of the document, the purpose of the instruction, and related references and forms. • Develop the work instructions. Ensure proper flow and use clear language and/or pictures. • Test the work instructions by allowing someone who is unfamiliar with the task, but has the level of knowledge for the target audience, perform the activity by using the work instructions [8].

2 Methods There are several types of operations in robotic cells in automotive industry. We are talking about input operation, output operation and everything between these two. Almost all input and output operations are carried out by human worker except ones with special operation conditions where collaborative robots must be used. Either way, in order to create fluent material flow with high quality performance and low error rate it is necessary to create work instruction for all operations in such robotics cells. Between input and output operations welding, riveting, bonding and other common operations in automotive industry can be found. These operations are carried out by robots, but it is necessary to create work instructions for these processes for controlling purposes as well. There are three main types of work instructions. Written work instructions are easier to create, but are not very suitable for operations with complicated tasks that can be found in complex production systems. Rather than that, visual work instructions take place. Not only for that, but according to the learning pyramid (refer with: Figure 1) a recipient remembers about twice as much of all information if these is conveyed through visual content. Visual work instructions may consist of pictures or even of videos [9]. According to all this work instructions can be divided in several groups.

2.3 Continues operations (Bonding, MIG/MAG welding) 2.4 Spot operations (Spot welding, riveting, Screwing) 2.5 Control operations 3. Dividing by method: 3.1 3.2 3.3 3.3

Use of text Use of images Use of videos Use of videos

2.1 Parameters set Creating of any type of work instructions cost lot of time and analyzing. In order to shorten this process, standard template for work instruction and standard flowchart of their modeling should be developed. This template is then used not only for creating new instructions but also for update existing ones. First step of developing template and flowchart is to choose method of work instructions. For our needs Saatys decision method is used. Several parameters are set: processing time, knowledge transferring efficiency, deployment difficulty. Weight of these parameters is set in figures (Refer with: Table 1–4).

1. Dividing by worker: 1.1 Human operation 1.2 Robotic operation 1.3 Combination – Human/collaborative robot operation 2. Dividing by technology: 2.1 Input and output operation 2.2 Manipulation

Figure 1: Learning eflciency pyramid [9]. Table 1: Parameters for decision making.

Parameters: A - Processing time B - Knowledge transferring eflciency C - Deployment diflculty D - Development diflculty


Creating Visual Work Instructions |

693

Figure 2: Comparison of time spend to process information from work instructions [8].

Table 2: Comparison of importance of each parameter.

α A B C D ∑︀

A 1.000 3.000 3.000 1.000 8.000

B 0.333 1.000 0.333 1.000 2.667

C 0.333 3.000 1.000 0.500 4.833

D 1.000 1.000 2.000 1.000 5.000

2.2 Variants comparison

Figure 3: Graph showing development of assembly time for each support tool [11].

Figure 4: 3D work instruction visualized on tablet [10].

Note that weights of coefficients are set mostly on experiences and worldwide available literature about work instructions.

According to Boer et al. [10] and Li et al. [11] the best option for short learning time are visual work instructions instead of written instructions. Figure 2 compares We can see few comparison researches on this topic on pictures below where measurements shows that it takes more time to process information from textual form of work instruction rather than from photographic or video-based instructions. (refer with: Figure 2, Figure 3). The JLX3D company [12] and Michalos et al. [13] take it even further into 3D animations and augmented reality, but it also said that it takes more time and investments to develop such instruction as it is at its infancy. (refer with: Figure 4). In Sandahl [9], Li et al. [11] and Hořejší [14] it is stated that if more senses are involved more is remembered. Thus, in order to create short term instruction with high knowledge transferring efficiency at least two senses should be used. For best solution augmented reality, or movie-based instructions should be created. According to Olson et al. [3] and Mourgues et al. [6] it takes several hours to produce a good, formal work instruction depending on the complexity of the instruction, availability and quality of the information, skills and knowledge of the person producing the instruction. While augmented reality and video-based instructions seems to be better for knowledge transferring efficient their developing and deployment is much longer and much more expensive. Based on this information comparison of all variants in accordance with all parameters are created (refer with: Table 5–12).


694 | M. Beluško et al. Table 3: Calculating standard weight of parameters.

α A B C D

A 0.125 0.375 0.375 0.125

B 0.125 0.375 0.125 0.375

C 0.069 0.621 0.207 0.103

Table 4: Final weights of parameters.

αi 0.130 0.393

B V1 V2 V3

0.277 0.201

Table 5: Comparison of variants in accordance with processing time.

V1 1.000 5.000 5.000 11.000

V2 0.200 1.000 1.000 2.200

V3 0.200 1.000 1.000 2.200

Table 6: Calculation of effectiveness in accordance with processing time.

A V1 V2 V3

V1 0.091 0.455 0.455

V2 0.091 0.455 0.455

V3 0.091 0.455 0.455 ∑︀

∑︀ 0.273 1.364 1.364 3.000

nij 0.091 0.455 0.455

Table 7: Comparison of variants in accordance with knowledge transferring eflciency.

B V1 V2 V3 ∑︀

V1 1.000 3.000 5.000 9.000

αi 0.130 0.393 0.277 0.201

0.519 1.571 1.107 0.803 4.000

Table 8: Calculation of effectiveness in accordance with knowledge transferring eflciency.


A V1 V2 V3 ∑︀

∑︀

D 0.200 0.200 0.400 0.200 ∑︀

V2 0.333 1.000 3.000 4.333

V3 0.200 0.333 1.000 1.533

2.3 Final decision Following the comparison equation (1) is used to find standardized weight of every variant for particular parameter (Table 13). Sums of these standardized weights are 0.313

V1 0.111 0.333 0.556

V2 0.077 0.231 0.692

V3 0.130 0.217 0.652 ∑︀

∑︀ 0.318 0.781 1.900 3.000

nij 0.106 0.260 0.633

Table 9: Comparison of variants in accordance with deployment diflculty.

C V1 V2 V3 ∑︀

V1 1.000 1.000 0.143 2.143

V2 1.000 1.000 0.143 2.143

V3 7.000 7.000 1.000 15.000

Table 10: Calculation of effectiveness in accordance with deployment diflculty.

C V1 V2 V3

V1 0.467 0.467 0.067

V2 0.467 0.467 0.067

V3 0.467 0.467 0.067 ∑︀

∑︀ 1.400 1.400 0.200 3.000

nij 0.467 0.467 0.067

Table 11: Comparison of variants in accordance with development diflculty.

D V1 V2 V3 ∑︀

V1 1.000 0.333 0.111 1.444

V2 3.000 1.000 0.143 4.143

V3 9.000 7.000 1.000 17.000

for first variant, 0.350 for second variant and 0.338 for the third variant. The highest one is optimal solution for our



695

needs and so it is used for next step of our research. U ij =

n ∑︁

α i · n ij

(1)

i=1

2.4 Developing flowchart For new standard picture-based work instructions, standard flowchart for designing has to be made. The destination for our work instructions are welding and riveting cellular robotics workplaces. In such workplaces mostly input/output operations and controlling operations are developed so a new step of this paper is to construct a flowchart for creating picture-based work instructions for input/output operations. Figure 7: First part of flowchart.

2.4.1 First assignments in flowchart The very first step for our flowchart is the definition of station and operation. In “Definition of operation” operation name and station number are set. Next steps are the definition of worker or workers, product or parts and used tools. Information of these are written in documentation parts of visual working instructions. For the needs of documentation, header and heeler are created in standard template so it is easy to find but don’t distract the worker from important information. In our case information that is important for the worker or team leader are written in header because it is logically the first thing that person sees after looking at instructions except pictures. To increase attention for this part red boundaries are created (refer with Figure 5). Table 12: Calculation of effectiveness in accordance with development diflculty.

D V1 V2 V3

V1 0.692 0.231 0.077

V2 0.724 0.241 0.034

V3 0.529 0.412 0.059 ∑︀

∑︀ 1.946 0.884 0.170 3.000

nij 0.649 0.295 0.057 Figure 8: Second part of a flowchart.

Figure 5: Developed header of work instructions.

Other additional information like date of creation or creator name of instructions are written into the heeler. Blue boundaries for heeler are used (Figure 6). In order to create quick readable instructions different approach for input/output operations and control opera-

Figure 6: Developed heeler of work instrucions.


696 | M. Beluško et al. Table 13: Final result of Saaty’s decision method.


αi 0.130 0.393 0.277 0.201

V1 - Use of text nij Uij 0.091 0.012 0.106 0.042

Variants V2 - Use of pictures nij Uij 0.455 0.059 0.260 0.102

V3 - Use of videos nij Uij 0.455 0.059 0.633 0.249

0.467 0.649 ∑︀

0.467 0.295 ∑︀

0.067 0.057 ∑︀

0.129 0.130 0.313

0.129 0.059 0.350

0.018 0.011 0.338

Figure 10: Instruction overwhelmed by pictures. Figure 9: Third part of a flowchart.

tions are created. First part of flowchart is created (refer with: Figure 7).

2.4.2 Input/output flowchart standardization Once all definitions are made and written in header and heeler, pictures of these product and tools are created. To create the ideal picture, part, product and tools should be placed in angle worker see them but not upright the part – axonometrically if possible. All these are scaled same way and putted into one page. In input/output operations there are several parts, manipulator (if there is any) and fixture. Very important is to place all the pictures of used tools or parts on right place in paper so it will be fast readable and understandable in very few seconds. If there are more than 4 pictures on one page, next one should be created with the same header and heeler with number of the page. Fixture is set in middle of the page while parts are set around it according to their position in fixture (part fixed

in a fixture on the right is places in a paper on the right from that fixture). Manipulator is set in free upper corner. If the page is overwhelmed by pictures work instruction is divided into two steps with roughly same number of tasks. The next page is created with same header and heeler. Picture of fixture on both pages are set on an opposite side than parts used in that steps. Same picture of manipulator is placed in free upper corner. Pictures of parts from the second step are moved to second page and placed around fixture according to their position in fixture. Dividing the process is repeated until an optimal solution is found or no more dividing is available. In that case larger scale of pictures is used. Second part of flowchart is created (refer with: Figure 8).

2.4.3 Final part of flowchart After picture placing is done, numbers referring to the order of operations are set next to the part connected with that particular operation and arrows referring to position-



697

Figure 11: Instruction after splitting into two steps.

Figure 12: Instruction without overwhelmed pictures.

Figure 13: Final visual work instruction.

ing of part in fixture are added. At the end, name of parts

and tools are added underneath to even fasten the information gaining process (refer with: Figure 9).


698 | M. Beluško et al.

3 Results According to the findings, Saatys decision and the constructed flowchart, one example of visual work instruction is developed. Figure 10 shows the first instruction attempt with overlapping pictures. After dividing the input operation into two steps one part is moved to another page where pictures of fixture and manipulator are copied. Even after division pictures are still covering through each other (refer with: Figure 11). Another division is not possible because only one part/operation left in every page. Changes of scale of pictures have to be made. After rescaling the pictures visual work instructions look clear (refer with: Figure 12) and the last part of the flowchart can be used. Last step is to number the particular operations, create arrows and add names underneath parts and tools. Final visual work instruction is made (refer with: Figure 13). Pictures of fixture, parts and manipulator are from external sources (Manipulator: “http://positech.com/wpcontent/uploads/2015/09/taurus_hydraulic_manipulator. png”, Fixture and parts: “http://www.computoolgauge. com/images/attributefixture3.jpg”) and name of them are fictional same as name of station.

4 Discussion In this paper is set new approach to solve work instruction development. According to findings picture-based work instructions are about the same effective in processing time as video-based instructions (Uij = 0.059) and their distribution is as easy as written instructions (Uij = 0.129). For fast development the is best option written instruction (Uij = 0.130) because there is no need of creating real pictures or videos from real production cells. For picture-based instructions it is Uij = 0.059 and for video-based instruction only Uij = 0.11. On the other hand, video-based instructions are best for knowledge transfer efficiency with Uij = 0.249 while written ones have only Uij = 0.042. Picture-based has Uij = 0.52. According to Saaty decision making optimal solution for fast knowledge transferring, fast worker processing time, easy developing and fast distribution of new work instructions are picture-based work instructions with Uij = 0.350. Following decision making, a standard flowchart for picture-based visual work instruction was constructed. Only input operation flowchart was developed because of diversity of operations in robotic cellular workstations.

Places for information in textual form were created on every page of work instruction so it is not necessary to browse through instructions to find them. The most important information was placed on very top of the page as it is the very first thing a worker reads, while additional information was placed at the bottom as it is the least distracting place for the worker. Every part, product and tool is placed between two textual areas and arranged to create intuitive instruction without need of text. Every part, product and tool is set to be at the same scale if possible. Name of parts, product and tools were added underneath them so worker can check whether his using right equipment or not. Arrows are added to emphasize intuitive section of instruction. At the end, an example of picture-based work instruction was created by following the steps in developed flowchart. The created example shows fixture, manipulator and parts with numbers that represent the order of their placement to specific places in fixture. Final work instruction is clear, intuitive, understandable, easy to produce or update and easy to remember.

5 Future work In the future, extended flowchart should be created incorporating control operations and their effectiveness should be tested and discus with workers in welding workplaces. Acknowledgement: This work is a part of these projects VEGA 1/0258/14, VEGA 1/0619/15, VEGA 1/0063/16, KEGA 006STU-4/2015, KEGA 018TUKE-4/2016.

References [1]

[2]

[3]

[4]

[5]

Molnár V., Boroška J., Dečmanová J., Mechanical properties of steel rope wires – quality test assurance, Acta Montan. Slovaca, 2010, 15(1), 23–30 Molnár V., SAP warehouse management system for a warehouse of auxiliary material in the selected company, Proceedings of the Carpathian Logist. Congr. CLC 2012 (7-9 November 2012, Jesenik, Czech Republic), TANGER LTD. Ostrava, 2012, 1–6 Olson M., Villeius E., Increased usage of standardized work instructions, Master thesis, Chalmers University of technology, Göteborg, Sweden, 2011 Greif M., Hamilton B., The visual factory – Building participation through shared information, Productivity press, Portland, Oregon, 1991 Peterson M., Standardization of process sheet information to support automated translation of assembly instructions and product-process coupling, Master thesis, Clemson University, South Carolina, USA, 2012



[6]

Mourgues C., Fisher M., Kunz J., Method to produce field instructions from product and process model for cast-in-place concrete operations, CIFE working paper, Stanford University, Stanford, 2008 [7] Olofsson O., How to follow instructions, WCM Consulting AB, [Online: “http://world-class-manufacturing.com/Lean/swi .html”] [8] ASQ Global, Work instructions, [Online: “http://asq.org/ser vice/body-of-knowledge/tools-work-instructions”] [9] Sandahl K., How to create visual work instructions, 2015, Gluu ApS, [Online: “https://www.gluu.biz/visual-work-instru ctions”] [10] Boer J., Blaga P., The implementation of new work instructions in te training of human resources in the process of paining in electrostatic field, Procedia Economics and Finance, 2015, 32, 514-519

699

[11] Li D., Mattsson S., Fast-Berglund Å., Åkerman M., Testing operator support tools for a global production strategy, Proceeding of the 6th CIRP Conference on Assembly Technologies and Systems (16-18 May 2016, Gothenburg, Sweden), Elsevier, 2016, 120-125 [12] JLX3D, Four steps to create modern work instructions, JLX3D White paper series, 2012, http://www.edstechnologies.com/ Mailer/Sep13/PF%202%20-%20Modern-Work-Instructions-JLX 3D.pdf [13] Michalos G., Karagiannis P., Makris S., Tokçalar Ö., Chryssolouris G., Augmented reality (AR) applications for supporting human-robot interactive cooperation, Procedia CIRP, 2016, 370375 [14] Hořejší P., Augmented reality system for virtual training of parts assembly, Proceedings of the 26th International DAAAM Symposium “Intelligent Manufacturing & Automation” (21-24 October 2015, Zadar, Croatia), 2016, Elsevier, 699-706
