Viewpoint

Viewpoint

Article Type: Viewpoint From: Industrial Robot: An International Journal, Volume 36, Issue 6

Geary V. SoskaCMfgE, Robotics, is the President of Robotics Applications Consulting

Industrial robots have proven themselves to be significant factors in the improvement of productivity, profitability, product quality and indeed, the competitive edge of many companies worldwide. But trying to implement them without a sound understanding of the technology and proper planning techniques can spell disaster for your company, especially if it is your first time.

Whether you are the executive who approves a budget to spend capital or a member of the technical staff charged with implementing your first application, there are certain things each of you needs to know before you take the big plunge. For example, executives need to know what industrial robot technology is and is not, what the capabilities and limitations of industrial robots are, what tasks are best and least suited for industrial robots, when they should or should not be applied, and confident that the budget they are about to approve has been thoroughly thought through. Those on the technical staff need to know all this plus more because you too are decision makers, only in a different sense. You also need to know how to put together a comprehensive budget that covers the total installed system commissioning costs, which means you need to prepare a very detailed installation and production commissioning plan. You need to know who to train and when and where the training is to take place. You need to understand when the robot's warranty takes effect and what it actually covers. You also need to know how to prepare a functional and performance specification. Never mind writing an equipment specification. You do not do it when you purchase a new car. In other words, you do not tell the automobile manufacturer how to build your car. You look over the product literature and make your decision based on your driving needs, vehicle style, and vehicle color. The same holds true with an industrial robot. All robot companies post their product specifications on their web sites. All you need to do is shop around and pick a model that best meets the needs of your application.

Let us take some time now to elaborate on some of the points I have raised.

Industrial robots are not a cure-all. Many times, companies turn to industrial robots to help resolve some kind of manufacturing process problem. This is one of the many formulas for failure. My advice is to identify the root cause or causes of your manufacturing process problem and correct it. By doing so, you may find out that you do not need to make an investment in an industrial robot, because if the process itself is bad, there is nothing the robot can do to correct it. Only people can do that.

When you prepare a budget for an industrial robot application think in terms of a total delivered, installed and commissioned system. If you fail to do this, you will probably find yourself going back to management to request additions funds. Not a position you want to find yourself in, and rest assured it is not one management wants to be in. Remember, by itself the robot is nothing. It is just a machine bolted to the factory floor.

On the subject of budgets, let me interject some food for thought. Most companies and government entities require competitive bids and will typically choose the lowest bidder. If you continue with this process for future applications you run the risk of mixing robots from different manufacturers in your facility. That is absolutely the last thing you want to do for several reasons. First, no two manufacturers' robots are the same. Their control systems are different, their operating systems differ, parts cannot be interchanged from one manufacturer's robot to another, and each manufacturer's robot is programmed differently via teach pendants or by using off-line techniques which are not universal. A Japanese-manufactured robot is not compatible with a Swedish-, German- or Italian-manufactured robot. If you mix robots in your facility, you face the following issues, all of which are cost drivers and can adversely impact worker health and safety.

With mixed robots in your facility, you face the issue of stocking spare parts for all of them, because the parts are not interchangeable. Since today's industrial robots are highly reliable and deliver up-times in the 99 percent range, people who operate, program and maintain them will have a tendency for forget and will need refresher training, especially when it comes time for reprogramming work. The question is, how much time and money do you want to devote to refresher training? I would not want to personally be in a situation where I had to know how to operate, repair or program three or four different manufacturers' robots. The confusion factor is just too great. One mistake and you can end up destroying equipment or perhaps having someone injured of even killed. You simply do not need that. I have worked in the automotive, agricultural/heavy equipment, aerospace and tire and rubber industries (in addition to the robotics industry) and I made it a policy to never mix robots in my factories. Instead, I found which I believed to be the best company with the best products and adopted them as a global company standard; similar to what the automotive and other industries did with programmable logic controllers. Once you standardize on one robot manufacturer's product, you will find out that they are more than willing to remain cost competitive. They want your repeat business. While I have my personal preference which has not changed in over 30 years, I will leave that decision up to you.

Another piece of food for thought is the evaluation of potential industrial robot applications. One thing I engrain into my students at the college where I teach part-time is that tasks that appear to be relatively straightforward for people are not necessarily as straightforward for an industrial robot. The best example I can give is tying a shoe lace. The average child can tie a shoe lace without even looking. When I am a presenter at robotic conferences I always pose this challenge to my audience: “Give me ten of the greatest minds in the world today, an unlimited budget and I'm willing to bet that after five years, they can't program an industrial robot to tie a shoe lace.” Too many variables and too much judgment coupled with real-time eye-hand coordination. Just some food for thought. I once called on a production director at a major jet engine manufacturing facility. He thought he had the perfect assembly operation for an industrial robot. What the operator was building was a section of the engine compressor. It was round, yet cone-shaped. The operator had this section on a turntable and would insert fiber bushings into holes in the section and then insert a metal pin into each bushing. Later, in the engine assembly process, titanium blades would be attached to the pins. This production director was convinced beyond the shadow of a doubt that this was his company's first and best application for an industrial robot. So, I asked permission to speak with the operator. I told her I noticed that as she was attempting to insert the fiber bushings, she would sometimes stop and apply what looked like a blue lubricant to the bushing and tap it in with a small plastic hammer. I asked her why she did this and she told me that sometimes the bushings were a little to snug to just be pushed in by hand. I also asked her why she sometimes tossed the bushings into a box on the floor. She told me that sometime the bushings were too small in diameter. I also noticed the same thing going on when she went to install the pins. I asked her why and once again she told me the pins were either too large or too small in diameter. I also asked her what the company did with the bushings and pins she tossed into the boxes. She told me that they were sent back to the manufacturer for credit. Do you see how much human judgment was involved with this process? Do you think an industrial robot would have done any better? Absolutely not! The robot would simply do what it was programmed to do and try to insert the bushings and pins whether they fit or not. Consider all the damage that could have occurred to a high-cost part.

After talking to the operator, I went back to the production director's office where he asked me what I thought about the application. I told him what he was dealing with were out of specification parts and that if his company would tighten up the specs for the bushings and pins, they could avoid the expense of an industrial robot which would have never worked. He took my advice and the company never had a problem with out of spec bushings or pins again. Like I said earlier, try to determine the root cause or causes in your manufacturing process problems and correct them. Then re-evaluate the application for robotics. Chances are you will not need one. You always want to be a penny wise, not a pound foolish.

Another point I would like to make is never try to go it alone with your first industrial robot application. You know more about the intimate idiosyncrasies of your manufacturing processes because they are yours. The robot manufacturers know far more about their products because they designed, built and improved upon them. Always work in partnership with your robot supplier and insist on totally integrated systems from them. Remember, your first robot application will set the stage for any future applications. If your first one fails, you can rest assured there will not be others down the road.

On last point related to training. The worse thing you can do is train your people in-house. The best way to do it is to have the training take place just after the acceptance run-off at the robot manufacturer's facility or at the system integrator's facility. There are three benefits to doing this. One, training is fresh in everyone's mind. Two, you do not have to shut down production operations to conduct the training in-house and three, those being trained will not be interrupted by factory emergencies. I have been in too many training situations in factories, especially with skilled trades people, where I have seen this happen. A quarter of the way into the day's training session; a maintenance supervisor will come to the class and need to borrow a few people (just for a little while). Two or three hours later when they come back, they have missed a lot. Then the company is unhappy when their people cannot perform. If you insist on doing your training in-house, then you will need some kind of manual operation in place running production until the training is completed.

Implementing your first robot application does not have to end in failure. Just pay heed to what I have said and you should enjoy success with your first and future applications.

About the author

Geary V. Soska has over 36 years experience in applied robotics and in teaching the subject at the college level. He has authored numerous technical papers and trade publication articles on the subject and was the 1991 recipient of the International Golden Robot Award.