Digital manufacturing for industrial robotic workcells.

Kittl, D. ; Stopper, M.

Abstract: The technical progress of the past few years in the field of industrial robotics enables a new integrated used, computer-aided planning and engineering method--the so-called virtual engineering--for the purpose of shortening development and start-up cycles, reducing risk as well as providing solid predictions for the functionality of industrial robotics workcell solutions.

This paper describes the essential changes from conventional engineering towards true virtual engineering which includes 3D-CAD and simulation techniques, testing of individual processes, robot offline-programming to the point of virtual start-up (altogether constituents of digital manufacturing for industrial robotic workcells) and analyzes and characterizes the adjustments which become necessary when transferring the virtual output (robot program) to the real system.

Key words: virtual engineering, start-up, realistic robot simulation, virtual robot technology

1. INTRODUCTION

Approximately thirty years in robotics is not a very long time but within this period since the first industrial robot was introduced to the market ground-breaking developments took place and an immense progress occurred. Especially in the past few years, where industrial robots became a common production resource also for medium and small companies and information technologies developed rapidly, building successful robot systems is no longer an issue of mechanical engineering and design alone. The programming as well as the understanding of the processes and motions these machines have to perform is the key to success (Adams, 2004). On account of this the main focus to achieve technical and financial advances is set on improved planning and engineering methods. Available key technologies for this purpose are

* high performance 3D-CAD

* realistic robot simulation based on true virtual robot technology

* high accurate system simulation

* true offline programming

* soft PLC and I/O simulation

and in combination with their integrated use and sufficient computing power a new process, the so-called virtual engineering, is created for planning and engineering industrial robot systems and solutions. The aim of this paper is to describe the differences between conventional and virtual engineering and to specify the therefore mandatory calibration process of the improved virtual engineering process with several practical changes and consequences pertaining to the transfer of the offline generated robot program to the real system.

2. PROBLEM STATEMENT

With the conventional planning process where the general procedure is a more or less sequential working method and where each phase is separately supported by computer-aided techniques any planning and engineering mistake leads to considerable delays and significant additional costs, for example when construction changes, maintenance repairs, reprogramming efforts or new resp. additional components become necessary. In consideration of the fact that there is no integrated scenario for the whole process available a process adaptation (using afore mentioned key technologies and alterations) is required.

[FIGURE 1 OMITTED]

Fig. 1. compares the conventional with the virtual engineering process developed by Kittl, 2005 and Stopper, 2005 with reference to earlier approaches by Peierl & Schlogl, 2004.

This process is now extended by the authors of this paper about the calibration process. Here it is important to allude that this extension becomes necessary by reason that the advantages of offline-programming face the difficulty of the calibration from the virtual output to the real system. A few practical adjustments have to be accomplished when transferring the offline-program to the real robot controller due to the fact that in the majority of cases the real world does not reflect the perfect virtual conditions in reference to assembly and measurements (think of a perfectly plane table top in virtual environments in contrast to the slightly uneven table top in reality).

3. APPLICATION AREA

Basically it is a wide area where virtual engineering is applicable and of avail. Considering solutions for industrial robotic workcells applicated by manufacturing and design teams, their contractors and supply chain, as well as service companies in many different industries the scope ranges from general activities like

* active customer involvement for solution finding

* functionality predictions already in an early phase

* workcell layout design and modeling

* analysis of the robot working area

* true offline programming

* robot path and movement optimization

* robot cycle time analysis and optimization

* developing and testing new procedures

* reducing risks and failures, thus optimizing costs

* etc.

to process specific activities such as

* automatic path generation along part outlines

* optimization of tool positioning

* component construction (e.g. grippers)

* and many more

Virtual engineering for industrial robotic workcells is solely developed and specified for the above mentioned category groups to optimize and modernize their own internal planning methods and on improving their processes during the engineering phase.

4. METHOD USED

To guarantee a smooth virtual engineering process flow the following software tools are absolutely necessary corresponding to the key technologies already mentioned in the introduction:

* 3D-CAD-Software: Essential to deal with CAD models (e.g. for construction, changes, format modification)

* Simulation Software: A program like ABB's Robot Studio which is based on true virtual robot technology and which has an RRS interface

* Offline Programming Software with 3D-graphics support (e.g. ABB's Robot Studio)

* System Simulation Software: A tool (e.g. Arena) to analyze the performance and the product flow of a system

* PLC and I/O Simulation Software (e.g. eM-PLC)

It is important that a data exchange between all these programs is possible to guarantee a completely integrated process and it is worth to mention that the support of virtual time (to obtain real cycle times in virtual environments) and interconnection technologies like OPC (to communicate between the different applications) is mandatory.

5. RESULTS

Surveying the extended process virtual engineering represents a much more compact and networked process where a couple of activities (simulation, manufacturing approval and offline-programming) can be executed in-house already in an early project phase. Fig. 1 shows the indicated differences in detail. The advanced process enables consequently the feasibility to provide solid predictions for the functionality of industrial robotic workcells and reduces cost and risk as well as shortens time-to-market (all represented by the [DELTA] sign in Fig. 1.).

This is one important aspect to use virtual engineering but taking away the uncertainties for the customer (oftentimes already during the offer phase) is necessary as well. This could be achieved by discussing the robot system created with 3DCAD with the customer and presenting video clips of the associated animations resp. simulation runs just before an investment is made for a project which possibly has not the value he expects.

While mechanical and electrical engineering has limited clarification possibilities at the conventional process, the supporting 3D-CAD layout design and diverse simulation analyzes as well as towards in-house displaced programming enable significant improvements applying the virtual engineering process. Given that there is normally just one technician on customer-site who carries out the whole programming at the conventional start-up situation, virtual engineering with in-house accomplished virtual start-up as a project milestone enables the use of knowledge of a greater number of start-up staff and allows better time scheduling because of the early project phase.

Upon completion of all in-house activities and after finishing the initial operation on customer site the transition of the robot program to the real system can be performed. As already mentioned in the problem statement it is mandatory to adjust the virtual output for using it in the real system. For this purpose the following practical steps have to be undertaken:

* work objects have to be measured on-site

* "Tool Center Points" (TCP's) have to be adjusted

* potentially important individual positions have to be (slightly) re-teached

* in case of a superior control (e.g. PLC) the robot program has to be suitably integrated

After all these calibration activities and the subsequent optimization phase the technical approval of the industrial robotic workcell can be arranged to complete the project in a briefer time and with a better performance as conventional engineering allows. In Fig. 2 the remarkable result of a use-case from the Austrian Robotics IT department of a huge Swedish company is shown where the virtually predicted cycle time is compared with the real process.

[FIGURE 2 OMITTED]

6. CONCLUSION AND FURTHER RESEARCH

The aim of this paper was to investigate and describe the differences between the in principle sequential conventional engineering process and the modernized, integrated virtual engineering process extended about the calibration.

Considering the economic values using virtual engineering significant cost savings, reduced risk, a shorter time-to-market and an increased performance can be obtained. Important milestones of the advanced process are the simulation completion--which allows visualizing and confirming the technical solution for the industrial robotic workcell before any commitment to hardware is made, the virtual start-up--thus the test run of the completed robot program in the virtual environment of a realistic robot simulation program, and the calibration process--where the virtual output is adjusted to the real conditions. Finally it can be said that digital manufacturing for industrial robotic workcells applying the virtual engineering process leads to economical, technical and operative benefits for planning and engineering robot system solutions.

7. REFERENCES

Adams, M. (2004). 'Start of production'= Termination point of the digital factory?, Proc. of SPS/IPC/DRIVES '04, pp. 245-253, ISBN 377-2367-09-7, Nurnberg, Germany

Kittl, D. (2005). Virtual Engineering in the field of industrial robotics, Diploma Thesis, FHW Vienna University of Applied Sciences, unpublished, Vienna, Austria

Peierl, H. & Schlogl W. (2004). Virtual start-up of mechanized stations as integral part of the digital factory, Proc. of SPS/IPC/DRIVES '04, pp. 123-131, ISBN 377-2367-09-7, Nurnberg, Germany

Stopper, M. (2005). Virtual Engineering for Industrial Robotic Work Cells, Proceedings of the 4th Asian Conference on Industrial Automation and Robotics (ACIAR'05), Paper ID: F-35, ISBN 974-8208-58-3, Bangkok, Thailand