Disassembling

Annual Reviews in Control

PERGAMON

Annual Reviewsin Control 23 (1999) 165-170

INTELLIGENT DISASSEMBLY OF ELECTRONIC EQUIPMENT

Bernd Kopaeek ~) and Peter Kopaeek.2)

1) Austrian Society for Systems Engineering and Automation Adlergasse 3/1, A-2700 Wiener Neustadt, Austria Tel.: +43/2622 27367, FAX: +43/2622 27367 - 22 E-mail: bernd, kopacek@ihrtnt, ihrt. tuwien.ac, at 2) Institute for Handling Devices and Robotics Vienna University of Technology Floragasse 7A, A-1040 Vienna, Austria Tel.: +43-1-5041835, FAX: +43-1-5041835-9 E-mail: kopacek@ihrt l. ihrt. tuwien, ac. at

Abstract: Fully or semi-automatized disassembly will gain importance in the nearest future especially for electr(on)ic waste. Concerning the recycling logistic - a new concept for this purpose based on ,,disassembly families" as well as a flexible, modular, intelligent disassembly cell is presented in this paper. Keywords: Recycling, Life-Cycle Engineering, Disassembly, Disassembly Cell

1.INTRODUCTION Manufacturers of domestic appliances, electromechanical and electronic products are facing one of their most serious challenges today: how to apply an integrated approach to business and product development to create environmentally compatible products, i.e. replace the traditional flow of products from manufacturer to landfill with a recycle and recover approach, because ex post emissions treatment and waste taxes are not sufficient solutions. A drastic change of scale is urgent. This can be obtained only by designing products more compatible with the environment. Furthermore, the world is facing an ever growing stream of electronics waste as a result both of the rapidly increasing number of new applications for electronics and of the accelerating pace of technological development and ever shorter service life. The tremendous increasing amount of electromechanical and eleetr(on)ic products, such as

computers, printers, telephones, household-machines and others, to be recycled (and also to be disassembled), makes it necessary to partially automate this disassembly process to increase the efficiency. The need for such semi-automated solutions can be estimated from the German electrical and electronic equipment market, where more than 2 million tons per year of electronic scrap are expected for the next decade. Within the EU the electronic waste has already reached staggering 10-12 million tons a year. In some branches there is a tremendous growth, like the German PC market, where 2,8 million PCs have been sold in 1994- worldwide the sale of PCs 1994 increased to 45,2 million (+25.2%). According to a recent study for the US market and estimated that if the current German pace of discarding 3-4 million computers per year continues, disposal costs would alone reach a staggering 200300 million ECU per year. Material and design selections are influenced by

1367-5788/99/$20 © 1999Publishedby ElsevierScienceLtd on behalfof the InternationalFederationof AutomaticControl.All rightsreserved. PII: S 1367-5788(99)00018-8

166

B. Kopacek, P. Kopacek /Annual Reviews in Control 23 (1999) 165-170

many factors such as engineering requirements. manufacturability, performance, environmental effects, costs and time. Such considerations are not mutually exclusive, but all decisions must be recognized as a compromise between different levels of reliability, performance maintainability and environmental friendliness in order to achieve minimum life time expenditures. Every product has environmental impacts which are determined by their design, adverse environmental impacts will be minimized or designed out. Decisions made at the design stage affect a product's impact on the production and recycling processes, the threats to human health and the environment and the characteristics of waste streams. Because of the complexity of electronic scrap recycling, various topics are addressed, like: • collection logistics, • disassembly, • reuse of usable parts, • recovery of precious and rare materials, • recycling of non-hazardous materials and • disposal of hazardous and toxic substances.

2. DISASSEMBLY (STATE OF THE ART) Until now a very high standard in the field of automation and robotics have been reached, but focused only on assembly. Few parts of electronic scrap are recycled after disassembling, however, the degree of automation is still very small - only some pilot or demonslration projects are realized mainly in research institutes. For the expected mass of products which will come back to recycling and disassembling companies in the future, the existing manual disassembly is totally insufficient. Especially parts of high quality products, i.e. those which contain precious metals, are disassembled in order to reuse some components. For a lot of products, especially those used in private households, the effort of manual work would not be worth it. Today most household products are shredded without any disassembly. In this case a separation of toxic components is not economically feasible and therefore not done at all.

are several end of life options existing, namely: • upgrade, • reuse, recondition, • re-manufacture, • resale, • recycling of materials and • disposal. Almost all of them require some form of disassembly. As already pointed out, the tremendous amount of electronic scrap deduce a need for automated solutions. Because of the particular characteristic and requirements of disassembling tasks, fully automated disassembling needs structures and methods for a semi-automated disassembling with both, use of manual and automated (e.g. robotized) workplaces to meet the requirements of a new life cycle strategy. At present, recycling and disassembling of PCs gains more and more importance. Until now, there are some pilots for automated disassembling of keyboards (e.g. Austrian Research Centre in Seibersdorf), monitors (e.g. Siemens, Germany) and printed circuit board (University Braunschweig) - but for there is no (semi-) automated solution for the PC itself. All existing pilot or laboratory systems for automated disassembling do not fulfil the demands for the future. Compared to the planned disassembly system, already existing concepts are very inflexible and only developed for a special task or product - e.g. few types of telephones, MiniDiscs or IBM Keyboards. If there are only minor changes to the product, these very inflexible systems must be adapted with very high costs. Hence for the industrial user such a system is only economical for large quantities of the same product. A further problem arises (in most cases) when used products are going to be disassembled. The characteristics (geometric form, dimensions,...) of these can differ enormously and a ,,non intelligent system" is not able to take them apart efficiently.

3. DISASSEMBLY CELLS Because of the tremendous increasing amount of electronic products to be recycled ( and also to be disassembled), makes it necessary to (partially) automate this separation process to increase the effort. However also for the disassembly processes high flexibility, high accuracy, vision sensors and lowcost will be necessary. The automation potential will be of the most important productivity factors for this new production process and becomes a new challenge for engineering.

As pointed out earlier ,,stilT' automatized disassembly in single purpose cells - only for one product (e.g. one type of PC's) - cannot be economically operated today. The number of devices or parts to be collected and concentrated on the place of the disassembly cell is usually to low for a two shift working of the cell. For example in the case of computer-keyboards of a distinct type, all keyboards disposed all over Europe per year could be disassembled in three month by this fully automatized cell.

Especially for electric and electronic products, there

Therefore forming of relevant product groups

B. Kopacek, P. Kopacek / Annual Reviews in Control 23 (1999) 165-170

167

Storage systems for tools and parts These modules can be chosen and combined according to the necessities of a cell for a distinct disassembly family'. The designing of such a cell can be divided in: Analyzing: Which parts are the most suitable ones for recycling? --, Determination of the operations for disassembly of the selected components. Decision: manually or automatically?

~mOing

--, Selection of the tools starting with the ,,easiest" operations.

Fig. 1. Parts of a disassembly cell. (disassembly families) will be the key factor for the whole system. Here the optimum for the characteristics of disassembly, design and recycling must be found. These so called ,,disassembly families" are groups of similar or different products that require nearly the same disassembly operations carried out with the same disassembly tools.

3.1 A modular system for intelligent, flexible disassembly cells The design of a disassembly cell is an important topic and is carried out today by concurrent engineering methods. Fig. 1 shows all different parts of a modular, flexible disassembly cell. According to the figure the main modules of such a cell are: --. Industrial robots or handling devices with special features like high accuracy, path- and force control.

Contrary to the assembly process for disassembly it's very important having no operations with single parts. Therefore the number of single operations is lower and as a consequence the costs of the equipment too. For the development of a modular, flexible disassembly cell some of the modules are available from assembly cells, but research on the following subjects is necessary: •

Conception of sensor guided robots for disassembly.

•

Evaluation of necessary tools and grippers which fit to the required separation techniques.

•

Conception of extremely flexible and modular gripping devices for disassembly.

•

Evaluation of possible implementation of methods of artificial intelligence in the cell control with a minimum of sottware costs and computing time.

•

Evaluation of best suiting vision system and product groups, considering present product categories and discernible future design trends.

Special gripping devices for a broad spectrum of parts with different geometries and dimensions --, Disassembly tools especially developed for robots Feeding systems disassembled

for

the

products

to

be

Transport systems - similar as for assembly cells Fixture systems for parts geometries and dimensions

with

•

Evaluation of interfaces for the integration of vision system to semi-automated cells, adding visual identification and physical measurement methods for the activation of the robot control: - pre-selection of parts, sample to measurement device, or vice versa, - sample or surface preparation. a

different

Manual disassembly stations

-

Intelligent control units able information from extended sensors

to

process •

(,,Low Cost'') Vision systems for part recognition Various sensors for force and moment limitations, position, distance, etc.

Cost estimation of adapting a vision system to the indentified needs of the product and the necessary robot manipulation considering the industrial environment.

168

B. Kopacek, P. Kopacek /Annual Reviews in Control 23 (1999) 165-170

3.2 Disassembly logistics

possible assembly operations (selected from a list of predefmed assembly icons). The semiautomatic linkage of the assembly icons with available tools, grippers, changing systems, etc. and an automatic refinement of the assembly operations generate the best value for the required cyclic time of the assembly cell. After the planning process has ended, the planning software produces a list of all required parts, the calculated cyclic time, the required costs and the maximal delivery time (Fig. 2).

Disassembling can only be done economically when the interfaces are satisfactory. That means there must be enough products at the disassembly factory and the separating technique must be able do realize its task. Hence the results of this task will help to build the system at the right place and to use appropriate separation techniques For the different products in the product groups the applicable separation technologies must be identified. Further on it must be def'med if the disassembling operations can be done in a destructive or nondestructive way. For each type of product a solution will be designed for the components that are not entering the automated system. Optimum problems concerning the transport of the scrap from collection to the recycling places - defined as ,,disassembly logistics".

As there are many similarities to the planning of assembly systems, the application area of ROBPLAN will be extended now to planning of robotized disassembly cells.

4.1 Required adaptations of the databases:

4. COMPUTER AIDED DESIGN OF MODULAR DISASSEMBLY CELLS In 1992, the Institute for Handling Devices and Robotics started to develop a planning tool for knowledge-based, semi-automated planning of robotized assembly cells - called ROBPLAN (Kopacek and Kronreif, 1996). The kernel of this planning system consists of four independent, relational databases: Symbol Database, Product Database, Component Database, and Planning Database. Using this database system, ROBPLAN allows a semi-automatic selection of the necessary assembly cell components by manual selection of

Product Database: To calculate the optimal disassembly depth and the cost/benefit ratio of the entire disassembling process, this database should include the values of the disassembled parts (concerning the different recycling processes - material recycling, recovering, etc.). For the complete description of the product, the database has to contain a detailed description of the part connections (how to solve them, possible tools for solving the connections). One key aspect for the ,disassembly process is to know the ,,constitution" (age, damage, etc.) of the product and expected complications caused by this conditions. Symbol Database: The pre-def'med assembly icons from the Symbol Database have to be extended with representative disassembly steps. Product Database

Component Database creation of the product database

~mbly

~ and

Symbol Database I selection of the

Planning Database

assembly llme and price calcukl~on

evalualion of the results

Fig. 2. Structure of the planning system ROBPLAN

B. Kopacek, P. Kopacek/Annual Reviews in Control 23 (1999) 165-170

Component Database: The description of available cell components has to be expanded with tools and systems necessary for disassembling. Especially, a detailed description of available sensor systems and their characteristic data has to be included to the Component D ~ b e s e . Planning Database: Equivalent to the planning process of an assembly cell, the results of the design phase are written into the Planning Database. As an extension for the disassembling process, the proceeds of the disassembling can be calculated (and compared to the cost of the cell) and included to this database in order to proof the efficiency of the planned cell.

169

5.1 The MiniDisc There are two different types of MiniDiscs in production: a playback-only and a recordable one. An assembled playback-only MiniDisc consists of the following parts: Upper Cartridge, Label, Disk, Clamping Plate, Shutter Lock, Shutter, and Lower Cartridge. The rec~rdable one has a Shutter on both sides. The Clamping plate is made of a special type of magnetic steel - the Label is made of paper. The Upper and Lower Cartridge as well as the Disc consist of Polycarbonat, the Shutter Lock and the Shutter of Ployoxymethylen.

4.2 Required adaptations of the planning process 5.2 The disassembly cell Beside some adaptations of the selection and input masks (causedby the extension of the databases), the planning process of the assembly process basically can be used without significant changes. For using the planning system ROBPLAN and for evaluating the results proposed by the system, the knowledge about optimal disassembling depth, strategies for disassembling etc. is indispensable.

5. DISASSEMBLING OF MINIDISCS - AN APPLICATION EXAMPLE Sony D A I ~ Austria is one of the largest producers of optical storage units - MiniDiscs as their main product. A s it is typical for every industrial production, some of the produced MiniDiscs do not satisfy the desired high quality standard. Due to the rising waste disposal costs and the high costs of human work, an automatic recycling of MiniDiscs was the key aspect of this research project.

The disassembly cell consists of two main components: a feeding system and the disassembly system itself. Furthermore there is the cell control unit, a transportation unit (see Fig. 3) between the two components, and sensors to control the operation. The feeding system takes the MiniDiscs from a container. A recognition or inspection of the particular MiniDisc is not necessary, because every disk is the same and there is no wear, pollution or damage. The transportation system is equipped with sensors to orientate and align the MiniDiscs. Afterwards the MiniDisc is taken to the disassembly system. There, it will be fixed and cracked with wedges from the side. Doing so, the Upper and the Lower Cartridge are separated. A vacuum gripper picks the Upper Cartridge (Fig. 4) and puts it into a special container. An optical sensor controls that the

Fig. 3. MiniDiscs on the transportation system

170

B. Kopacek, P. Kopacek /Annual Reviews in Control 23 (1999) 165-170

Fig.4. Removing the Upper Cartridge Upper Cartridge reaches its destination. At the next stations of the system the Clamping Plate, the Shutter Lock, and finally the Shutter are removed with special tools. The proper course of these operations are also controlled by sensors. Every part is given into a specific container, where they are stored for further processes. Every 4 seconds one MiniDisc is disassembled by this system.

6. CONCLUSIONS Disassembly automation especially for electr(on)ic devices is absolutely necessary worldwide in the nearest future because of the dramatically increasing amount of electr(on)ic scrap.

modules for such cells or in their design and development. They can use such cells for an efficient disassembly of electr(on)ic parts. Such modular, intelligent, flexible disassembly cells ensure high flexibility necessary for the factory of the future. The planning and development process of such cells is a time-consuming process today. Therefore a computer aided planning system as an efficient tool for reducing the planning time DEROBPLAN - is under development. The system supports the planner and shortens the planning time.

REFERENCES

Today only some pilot projects in form of fully automated single purpose solutions - and only for one distinct type of one product - are realized. Main problem is the collection and transportation of a sufficient amount of these products to one place in one area.

Gschwendtner G., P. Kopacek (1995). Design for Disassembly. In: Preprints of the 4th Intern.

Therefore it is necessary to disassemble more than one product in an automated or semi-automated cell. For this purpose products with similar disassembly operations are collected in ,,disassembly families" families which are disassembled in modular cells, developed by means of computer aided planning systems.

Proceedings of the 8th Symposium on Information Control Problems in Manufacturing, Beijing,

This offers - especially for SME's - two possibilities: • They can be involved in the production of

Workshop Robotics in Alpe-Adria RAA "95, PSrtschach, Vol.l, p. 79-82.

Region,

Gschwendtner G., A. Frotsclmig, P. Kopacek (1995). Concurrent Engineering for Disassembling. In:

p.79-84. Kopacek P., G. Kronreif (1996). Robplan: Semiautomatic planning of robotized assembly cells. In:

Proceedings of the 27th International ~ymposium on Industrial Robots, Milan, p. 385-390.