CIM in 80's and now
'CIM' or Computer Integrated Manufacturing was coined in early 80's. Big picture was to integrate product design to various manufacturing processes with software and computers and using standardized protocols when transferring data between computers (MAP, Manufacturing Automation Protocol). As all hypes, the acronyms tend to come first, and after a while comes the break-through technology.
When it comes to robotics one application of CIM was robot simulation and offline programming (OLP): you take the information of a CAD design, based on CAD information you do a robot program offline in a computer, and check the tool paths for collisions and reachability by applying robot simulation. Finally you translate a program to a specific robot code and download that program to a robot on the shop floor.
What was the OLP technology looking like in early 80's. CAD models were mostly of wire frame (e.g. Calma CAD), computers were huge UNIX machines (HP, IBM, Silicon Graphics) and OLP software itself was based of wire frame models. Machines were also slow and software really hard to use. Real-time collision detection was out of scope. Industrial robots on the other hand were not well suited for offline programming: they typically had only five axis which made the programming and simulation offline even harder.
Number of OLP vendors was quite limited. Here are few software brands from the 80's: Silma, OLP from McDonnell Douglas (don’t remember the name), Grasp (BYG), RobCad (Tecnomatix), IGRIP (Deneb), even Finnish NOKIA Corporation made their own software. OLP software could easily cost 60 000$. Regardless the relative big investment cost, some buyers could calculate short payback times of an investment if robot teaching was quite regular causing robot down time to become significant.
Early adopters in industry were big automotive OEMs, like GM and Ford and aerospace industry, like McDonnell Douglas and Northrop Grumman. In Nordics the first pioneers were Tampella (Valmet), Bronto Skylift (Federal Signal company), Valmet Automotive, Volvo Aero, Volvo CAR and SAAB. The software was used for arc welding, sealing car bodies, spot welding and coating.
One significant break-through was OpenGL and PCs. Silicon Graphics (remember the animations of Jurassic park) had licensed their Graphics Library (GL) code which became Open GL. That made a PC powerful enough for fairly intensive robot simulations and handling 3D solid models. Now also medium size industries started to invest in OLP. Software prices had come down to half of the prices of 80's. Robot vendors had also started to evaluate software and some were bringing their own OLP brands. E.g. ABB's first 3D offline programming software was based on Deneb's (Detroit, MI) IGRIP software. The software was a stripped version of IGRIP and branded by ABB.
Today we can really talk about 'CIM'. OLP software is available with reasonably low prices and almost all robot vendors have their own software. All manufacturing companies are using 3D CAD. There are various so called generic OLP software brands which can be used with different robot brands. Some software are very powerful when utilizing CAD geometry and topology. In early days offline programming a body frame of a harvester could easily take 40-60 hours depending on the software. Now when using the most advanced software the same job could take two to four hours with all reachability studies and collision detection included.
Still robots in many companies are programmed by teaching even if that leads to low utilization rates and low overall equipment effectiveness (OEE). OLP technology of today makes 'CIM' to be reality in robotics. Investment in OLP will pay back in short time. Besides leading to shorter lead times OLP has also an impact which is not discussed that often: it improves quality and especially keeps the quality constant.
Heikki Aalto, 17.12.2014