Greg: I am always looking for people with interesting and enlightening application experience to feature in this column. I was fortunate to have Jim Cahill, Emerson social media manager, point me to Francisco Campa, director of manufacturing operations for Mohawk Home, a division of Mohawk Industries. Even when the economy was hurting, Mohawk Home was breaking sales records. At Mohawk, Francisco has demonstrated considerable success in applying technologies such as robots in manufacturing operations for home products that are outside the areas of expertise and the process industry applications of my career. One of the great benefits of this column is the opportunity to develop an understanding of innovative automation accomplishments that are the result of extensive effort and expertise in an hour of conversation. This knowledge and recognition of achievement leads to the advancement of the automation profession. This is my main motivation.
Stan: Francisco, as a previous configuration engineer and director of technology addresses both the technical and business aspects important for maximizing benefits, what is your general approach?
Francisco: Everybody agrees to automate, but how do we pay for it? I develop and justify automation based on whether it makes good business sense. I seek the business case in terms of contributions to the bottom line. I develop the scope, architecture and technology. During my career I have done tens of millions of dollars of highly successful automation projects using this approach.
Greg: What was your most recent project?
Francisco: I was looking at using robotics to help bring manufacturing back from overseas. In this particular plant for reprocessing rubber tires into floor mats, the labor costs and human safety risks were excessive. The process is very labor-intensive and hazardous with many recordable incidents a year. I took the difficult and dangerous operations out of the hands of the operators. The result during this first year of operation is no recordable incidents, a 70% reduction in labor, 60% increase in productivity and an 80% improvement in quality. Consequently, we were able to reduce the number of production machines required, which reduced capital costs.
Stan: What are some of the details of the manufacturing process?
Francisco: In the key part of the process, molds are filled with recycled crumb rubber and then put into a high-temperature press. All of this was previously done manually. A robot now handles the molds by tending the press. The robot grabs a completed mold and empties the product onto a conveyor, then presents it to a filling station to be reloaded with material. After a set number of molds are ready, the robot puts them into the press. The down time between pressings was reduced by 85%, significantly increasing equipment utilization.
Our original project only considered a robot to just grab finished goods and put them on a pallet, but very quickly we realized that most of labor, quality opportunities and safety concerns were in the main molding department.
Past attempts to automate this manufacturing process showed limited success; I envisioned a robot as providing a flexible solution that could help us achieve our goals.
Greg: How did it work out?
Francisco: The robotics portion of the project was easy, but everything else was problematic. Robots have very high uptime numbers and are very reliable; tasks are easily programmable and for the most part, robots will carry out their tasks with no issues. However, the equipment around the robot is not as reliable and will show lower performance numbers.
Our initial productivity was only 10% of expected; by successively addressing these support processes we increased productivity to 60%, then 100%, and finally to the incredible level it is today.
Stan: Even a great solution can be for naught unless the whole process is addressed. Problems upstream or downstream can result in failure. How did you solve these problems?
Francisco: It is important to evaluate the complete value stream for the manufacturing process and not only that portion to be automated. Prior and subsequent process will have a significant effect on the automation's success and the performance. In our case, we included all supporting processes including press retooling and raw material properties. By doing this, we also were able to almost completely eliminate the downtime required for product and process changes and improved product performance while allowing for increased throughput.
The crumb rubber, once mixed and prepared to be used in a mold, has a very short shelf life. You only have so many minutes to get the mix into the mold and into a press. We also required the use of supplemental additives to aid in product extraction from inside the molds, which, if not properly used, could create quality issues downstream and buildup around equipment.
We benchmarked and hired consultants to assist on these supporting processes, but the early suggestions were not very successful. Everybody was focused on the already existing solutions, and thought we simply had a task of selecting the appropriate one. Depending on your goals, this approach might not provide the right solution to the challenge and will limit the project's success. I asked fundamental questions about what was happening in a different way. By doing this, we found game-changing solutions to the problem.
Everybody was focused on robotics, since the technology is glamorous, and would very easily lose focus on other contributing processes and technologies involved in this project.
We had six different companies working on their respective aspects of the project: robotics, presses, mold design, raw material handling, wet material handling and overall automation.
Greg: What are the lessons?
Francisco: You need a complete scope for the automation project to address every step in the process. If we had stayed with the original narrow scope, some processes would have been left manual and unimproved, and we would have had little or no benefits.
You don't want something complicated or fancy, you want the simple, elegant solution that safely addresses the heart of the problems while meeting economic objectives. The optimum solution in terms of using robots was outside the fence.
Overly complicated automation and projects are counterproductive. As a customer, you may think you want all the bells and whistles, and the contractor will be pleased to deliver. You end up paying for the extras not only in upfront costs, but more importantly, in terms of complexity and maintainability. You need to stick with the essentials and allow for improvements as you gain experience with the system. A significant part of asking for what you really need is knowledge gained from previous experience and training.
We failed to train people early enough in the project in new technologies. When we got into training and realized what was under the hood, we were too late to realize the overcomplicated systems and to fix some of the limitations. We needed to train everyone involved before we even started on the project in order to be ready to ask the right questions and challenge the implementation process.
There must be a balance as to how far you go with automation. Above all, you don't want to automate a bad process. Automation should be considered as a tool from our continuous improvement toolbox, and automation projects should be based on process optimization rather than design a process around the automation technology.
Stan: We have had some projects that focused excessively on the labor savings resulting in the business case being a head count reduction. How do you avoid this negative approach that discredits operations?
Francisco: We try not to automate with the sole intent of reducing labor costs. When we discuss the financials, we are focused on productivity, quality and safety. The increase in productivity can be taken as an increased production rate, a decrease in manufacturing costs, including labor per unit produced, and an increase in plant viability. You must define metrics that will follow the production demand and consider the project returns at all levels.
There are many cases where you cannot eliminate a human. We can help the human be more proficient and efficient, and take away risky, repetitive operations. The complexity to eliminate a person increases exponentially and is impractical in many cases. This can prevent you from being able to justify automation or cause you to aim for goals that are unachievable.
Greg: The concept of a "lights-out plant" was inherently flawed. Some processes can be run remotely, but inevitably there is a human involved somewhere.
Stan: What other words of wisdom can you share with us?
Francisco: Automation projects justification is not an easy challenge. You must match the technologies available to the current business case. This involves identifying your current manufacturing processes' weaknesses and business threats; through automation and process improvements those weaknesses will become strengths that will turn business threats into new opportunities.
The second question is always around why now and not later. As an automation engineer you must understand the complete business and be able to match the two, allowing you to create a project case or automation need. From here, a plan can be developed showing a clear business benefit, a plan that also includes a scenario of what would happen if it is not implemented. Justification of a project must include both the benefits of doing it and the risks of not doing it.
Projects need to be presented from the point of view of the business and not the technology itself. Ask the question, "What's in it for them?" There is a big difference between "features" and "benefits;" remember that we always buy "benefits" and only after everybody likes the benefits, you will be allowed to invest in the features.
In our particular case for robotics implementation, we went beyond the labor reduction savings. We included opportunities to reduce waste, reduce variability, improve on safety, improve on compliance and to allow us to stay in business in a competitive market.
Overdesigning and overcomplicating projects costs companies money and implementation headaches. Identifying how far the implementation should go, and what kind of technologies to use, should be on top of your list. A key concept that can help you define these aspects revolves around the project type at hand. I usually think of projects as a strategy that will either solve a problem or simply improve an outcome.
Projects designed to solve problems are normally the ones that use a well-defined architecture and technology. Automation can normally be pushed to the limits, and game-changing solutions are developed.
On the other hand, projects designed to improve outcomes are those that we do with the clear understanding that there is no ultimate solution; the problems will not be completely solved, but their effects will be minimized or dealt with. On this type of project, we look for a limited implementation of technology and match that to human interaction, allowing us to improve on a situation or status quo.
There are potentially many more applications of robots in manufacturing and at home. See the online version for Greg's tongue-in-cheek "Top 10 Tempting Uses of Robots at Home."
"Greg's Top 10 Tempting Uses of Robots at Home"
(10) Totally perplex solicitors on the phone
(9) Walk dog in all types of weather
(8) Clean out litter box
(7) Supportively listen to spouse
(6) Entertain mother-in-law
(5) Make small talk at parties
(4) Play poker with friends
(3) Buy presents for relatives
(2) Do your income tax returns
(1) Sit in on your performance review