Mechatronics Part I: Motion Control’s Next Top Model
| By: Kristin Lewotsky, Contributing Editor
By using intensive modeling and simulation to discover problems in the design stage, OEMs can improve machine performance and save big.
A decade or two ago, motion control meant augmenting a system with a few servo axes. Today, hundred-foot-long packaging lines formed of multiple machines with dozens of axes are not at all uncommon. As the sophistication rises, the complexity of design grows exponentially - as does the likelihood of unanticipated interactions. Depending on the compliance and the masses involved, for example, damaging resonances and vibration modes can emerge when a machine is actually up and running under production conditions. Unfortunately, by that point, fixing the problem can be costly, if not impossible. To avoid developing a design that looks great in the computer-aided design (CAD) program but doesn’t perform, machine builders are increasingly turning to mechatronics
Once dismissed as a buzzword in some quarters, mechatronics refers to an integrated design approach that leverages multiple levels of modeling and simulation to optimize machine development and performance. Depending on whom you ask, the discipline emerged as early as the 1960s, during the Apollo moon-landing program. It wasn’t until the 1990s, however, that mechatronics truly matured into a multidisciplinary systems engineering approach.
At its most basic, mechatronics – or model-based design -- involves modeling the mechanical, electrical, controls, and even real-time software elements of a system (see figure 1). The idea is to evaluate the performance and understand the interactions to find a design that will best meet specifications. The trend toward mechatronics is driven by practicality: More efficient design means fewer delays and faster time to market, even in these times of limited resources, “Integration in the machine concept really requires an integrated approach in the design,” says Razvan Panaitescu, Engineering Manager, Mechatronic Support, Siemens Industry Inc. (Norcross, Georgia). “If you want to have a machine that performs from all perspectives correctly, you need to approach it accordingly from the very early stages.”
“It's very much about innovation,” says, Kevin Craig, Robert C. Greenheck Chair in Engineering Design at Marquette University (Milwaukee, Wisconsin). He identifies four elements to the process of innovating successfully (see figure 2). First, a set of feasible solutions must be developed based on mathematics and engineering principles. Next, those feasible solutions have to be adapted to the user with human-centric design. “You have to know what is desirable, you have to solve the right problem,” he says. “And then, of course, there is the business aspect. It's feasible but is it viable? Is it sustainable, energy-efficient, environmentally friendly?” The fourth and final element is complexity management. A design may be feasible, human-centric, and viable, but if it is too complex, it remains unusable no matter how desirable it is. Mechatronics provides a tool to achieve all four objectives. “These four elements have to be part of your solution and that's where innovation happens,” Craig adds.
Perhaps because machines used to be primarily mechanical, the established development paradigm has been that the mechanical engineers generate the machine design, pass their drawings along to the electrical engineers to add in the relevant drives and motors, then to the controls engineers to add in the logic, etc. All too often, companies following this serial approach complete the design and build the first prototype, only to discover serious problems in the system.
In today’s global marketplace, extreme performance demands, intensified by the pressures of a struggling economy, are pushing more OEMs to consider using mechatronics. At a time when machine builders are trying to eke out the greatest possible value for their engineering dollar, model-based design provides a means to optimize performance, improve return on investment (ROI), and manage risk.
The kinematics of motion are based on position, speed, and acceleration. We all learned in basic physics that F=ma, and yet it is not as simple as that. Perhaps a motor can move a 5 kg mass 1 m in 5 s, but what happens when that motion has to take place in 1 s? Do you get ringing, and does that compromise positioning? If so, perhaps the controls system can compensate. If the mass is doubled or the movement repeated at 60 Hz, though, oscillations and resonances might appear that are as unmanageable as they are unexpected. This is a case in which mechatronic modeling can both identify and isolate problems.
“You have to go deeper and analyze the construction of the machine and how the motor and the mechanical parts are integrated,” says Panaitescu. “Maybe the couplings are not stiff enough, or the pinions and gearboxes are too weak to withstand the high forces or high torques that are developed. If you analyze the machine more deeply, you may observe modes of vibration, eigen frequencies that are specific to that machine. You can either try to fix the mechanical issue or, if that is not possible, to absorb those unwanted frequencies through external means or other techniques like implementing controller filters.”
Before a problem can be addressed in the design phase, however, it needs to be discovered. The modeling process might start with a simple lumped-mass model to assess basic motion of the masses in the power train, assuming the various parts are rigid enough to be treated as a single piece. If the machine is not sufficiently rigid, finite-element analysis can be used to provide a better assessment of more compliant systems. It is important to note that a number of sophisticated software tools have been developed to assist with mechatronic modeling (see figure 3). Packages can import data from CAD programs or allow engineers to combine their finite-element model with the mathematical model of the controls to produce a complete virtual simulation of the machine. Mechatronic modeling can not only detect hidden problems it can assist the design team in choosing the best hardware for each motion task, so that the motors are powerful enough but not oversized.
It sounds like a long, complex operation but it doesn’t have to be. “Once you have drawn the basic components that are going to be included in that mechanism, the software tools nowadays allow you to do an analysis [very quickly],” says Panaitescu. “Software can import a 3-D design and produce a multi-body or a finite-element model. [It would take] approximately 6 to 8 hrs, maybe a day, to add all the conditions and constraints into the model. The simulation itself and the analysis would take another 1 to 2 hrs. You can bring this whole process into the regular design process and end up with a thorough analysis of what happens when this machine is put into dynamic forces.”
Handled properly, adding mechatronic modeling to the design process doesn’t have to be overly time-consuming. “Depending on how you realize this integration between the different design groups, you could end up not adding any additional work,” says Christian Fritz, Product Manager for Motion and Mechatronics at National Instruments (Austin, Texas). “It takes time but it also takes time off of the development process - rebuilding physical prototypes, testing, things like that. You just shift some of the implementation work to an earlier design.”
“It’s actually much more than a zero sum game,” says Craig. “The long-term savings in the cost of fixing problem and improving performance is enormous. You now understand what you have built. It has not been a process of design, build, test -- and pray that it either works or you can fix it if it doesn’t.”
There is an adage among machine builders that if an error caught in the design phase costs $100 to correct, it will cost $1000 if detected in the machine building phase. If no one notices it until the machine is installed at the customer's location, the price of fixing it rises to $10,000, which includes not only materials but airfare and engineering hours. It does not, however, include the cost of a dissatisfied customer who may not make a purchase the next time around. “People say they don’t have the time to do [mechatronics design],” says Craig. “My answer is you don't have the time not to. If you don't do it, you have no competitive advantage because you can't innovate.”
Savings can arise not just in minimization of hardware costs or early detection of errors, but in reduced cost of ownership. John Pritchard, Global Product Marketing Manager at Rockwell Automation (Milwaukee, Wisconsin) points to a customer who upgraded their 35-piece-per-minute packaging machine to handle 50 pieces per minute. "They called us back about six months later, around the time gas had gone to four dollars a gallon, and said, ‘Hey, can a mechatronic design approach reduce energy consumption as well?’” he recounts. “We looked to the model and we found a 64% energy reduction [from the original design]. It was there all along but we never noticed it.”
Instead of focusing purely on performance, mechatronics simulations can be recast to optimize around energy efficiency. In some cases, the energy savings may be minimal, or offset by the increased hardware cost for the design. If the energy savings over the course of a year are on the order of thousands of dollars, though, and the cost differential for the altered machine is only hundreds, it could very well be in the client’s best interest. Granted, purchase approval tends to hinge on return on investment (ROI), but long term reduced cost of ownership can be a powerful argument.
“A machine builder once said to me, ‘I think of myself as being in the design risk management business,’” says Pritchard. “He said, ’There are two basic risks in my business: One risk is if we don't innovate, we will be uncompetitive; the other is that every time we redesign something we break the number one rule which is if it's not broke don't fix it.’” It neatly summarizes the perpetual dilemma of the OEM.
When it comes to holding on to market share, good enough is no longer, well, good enough. At the same time, changes -- especially complete redesigns - can create nasty surprises and expensive headaches. For many years, machine builders resigned themselves to the notion that new designs would involve repeated failure en route to developing a machine that worked. The question was not whether a prototype would crash but how badly.
Merging electrical and mechanical systems can give rise to unexpected interactions that can either be discovered the hard way or through mechatronics. Consider a team developing a lower cost robotic arm for paint-spray applications. The challenge is to optimize mechanical stiffness, motor sizing, and control action. It is critical to ensure that the control system will give the arm stability at all points of reach for various possible loads to ensure paint is going where the customer wants it applied. “Using simulation prior to hardware prototyping to test the mechanical, electrical, controls, and software interaction can be a more comprehensive and less expensive way to discover that your control strategy is not correct,” says Tony Lennon, Industry Manager for Industrial Automation at the MathWorks (Natick, Massachusetts), “and that you really need feed-forward control in addition to PID control to account for some operating region instability.”
Granted, the simulation process required to discover those issues demands a certain effort, with associated cost, but in the long run, the cost of failure can be higher. “Say you install a design variation of a robot at the plant of a very highly valued customer who specified a certain throughput and you find out that your robot has to be slowed down because it cannot give repeatable performance at the speed to which you agreed,” says Lennon. “The customer’s location is not where you want to find that problem.”
One approach is to stick to tried and tested designs but that carries peril of its own. The alternative is to use model-based design as a tool to mitigate risk. “When I talk to machine builders who have adopted a mechatronic design approach, generally their innovation has accelerated many times over a machine builder who sticks to what they know," says Pritchard. "The whole question of how you innovate and mitigate your risk at the same time, that's a huge area for mechatronics design.”
The use of mechatronics design principles does not mean that design teams can eliminate prototypes. What it does mean is that the engineering team can design and optimize the entire machine before any metal is ever cut (see figure 4). By the time they build their prototype and do live testing, they are validating sound design, versus using the prototypes to find flaws. “Customers will admit that they have to get out of this mode of using prototypes as a way to find mistakes in their design,” says Lennon. An integrated mechatronic development approach lowers the risk of the entire design process and ensures customer satisfaction.
Especially as the economy recovers, OEMs need to field quality product, reliably and rapidly. Model-based design provides a powerful tool for achieving that goal. It doesn’t matter if a company has a wall of cabinets holding machine designs; the old way simply isn’t effective any more. “Rather than ‘plug and pray,’” Pritchard says, “what about spending some time using a mechatronic design approach to help you have a high degree of confidence that when the thing is finally built, it will be exactly what you want?”
In Part II of this article, we will discuss implementation of mechatronic and the barriers to success.