Motion Control Resources
- This technical feature is filed under:
Motion Trends to Watch in 2016
by Kristin Lewotsky, Contributing Editor
Motion Control & Motor Association Posted 12/16/2015
Only a few years ago, the proliferation of intelligent components dominated motion news. Today, it’s all about connectivity and the gains to be had from information. That’s important technology to watch, given that the market for connectable devices used in manufacturing production is slated to exceed $280 billion by 2019. Manufacturers of both consumer and industrial products that increase their market share in the coming decade will be those who gain a competitive advantage by adopting new technologies like big data, the Industrial Internet of things (IIoT) and smart energy monitoring for transparent factory operations.
There was a time when safety began and ended with relays. More recently, drives offering functionality like Safe Torque Off were considered sufficient. Today, focus of safety is broadening beyond protecting operators and equipment to encompass protecting performance.
The idea of applying functionality like Safe Limited Speed to use the power of the machine to clear a jam, for example, isn’t exactly new. Today, some organizations are going beyond that to use Safe Speed and Safe Direction to make a variety of operations less tedious and more productive. Regulations in market niches such as food and beverage mandate regular cleaning. With old-style safety equipment, an operator might have to stop the line to clean the exposed portion of a roller, step out of the enclosure to jog the machine, stop it, and step back in for another incremental cleaning. Because of the frequent delays, the process could take hours. Modern safety equipment can turn this into a continuous operation.
One food and beverage manufacturer was able to reduce a six-hour cleaning process to less than two hours - the equivalent of half a shift - by using these techniques. Elsewhere, an automotive supplier realized a 10% reduction in cycle time simply by retracting a part at full speed using Safe Direction after manipulation by an operator rather than at fractional speed using Safe Limited Speed.
These are just some examples of the kinds of productivity improvements available to machine builders, integrators and end-users who fully embrace the possibilities enabled by modern safety technology and standards.
Sensors and Predictive Maintenance
Today’s smart components leverage a significant number of sensors to enable comprehensive condition monitoring. Examples include proximity sensors, position or velocity sensors, temperature sensors, pressure sensors, accelerometers, and current sensors. If that sounds like a lot, it is - analysts expect the sensor market to exceed $115 billion by 2019, for a CAGR of 7.3%.[Reference 2] Condition information delivered with this density enables many things, first among them predictive maintenance.
Maintenance has evolved over the years from breakdown maintenance (making repairs after failures) to preventative maintenance (preempting faults by replacing parts and performing maintenance on a model-driven schedule). Depending on the product being manufactured, downtime can amount to hundreds of thousands of dollars per hour. At the same time, preventive maintenance ties up capital in spares inventory and consumes staff hours potentially replacing equipment that is perfectly fine. Comprehensive sensor data allows organizations to implement predictive maintenance, in which sensor data allows the system to detect when equipment is beginning to wear.
The density of sensors in today’s factory can alert maintenance to a variety of problems. An increase in temperature could warn of lubricant breakdown in a gearbox, for example. A characteristic vibration spike can indicate a bearing that’s ready to fail. Instead of courting unexpected failure, maintenance teams get early warning so that a subpar component can be replaced as needed during scheduled maintenance periods rather than during a crisis. Coupled to interface equipment that extracts this data and sends it to maintenance for analysis, sensors can provide significant savings.
Adaptive Self-Tuning Drives
In a perfect world, an axis would position the load as commanded. In reality, machine compliance prevents that from happening. Driven by the ratio of load inertia to motor inertia plus shaft wind up, each axis introduces characteristic resonant frequencies (see Figure 1). If the load and motor are moving out of phase with one another, the motion of the load becomes sluggish (anti-resonance). When they move in phase with one another, it dramatically amplifies the response (resonance). In either case, motion is no longer predictable. The spikes effectively limit the usable bandwidth of the system. Tuning the servo drive can help address this issue.
There was a time when commissioning a servo axis required a skilled engineer to painstakingly tune the proportional, integral, and derivative elements of the drive to optimize performance. Depending on the engineer’s experience and the complexity of the response, this could take anywhere from an hours to days. The emergence of autotuning drives simplified the process.
The initial autotuning routines only corrected device performance to a first approximation of what was appropriate. That significantly simplified the process but it’s still required manual tweaking. For 100-axis printing machine, that tweaking could add up to a lot of time. Today’s autotuning drives not only work better than the autotuning drives of yesterday, they bring important new benefits.
Autotuning machines can be divided between proactive and reactive. In proactive tuning, the drive exercises the axis, typically through a canned subroutine, to identify problems that cause excess of acceleration or ringing. With these results, the system filters the command signal to proactively removes any frequencies that excite resonances, bringing the load to the desired position more efficiently.
Reactive autotuning allows the drive to respond to machine-based vibrations by developing a notch filter to remove the resonance spikes. In both tech commissioning, they can significantly improve performance. The problem arises with the type of normal wear and tear that we discussed above. Maybe a nut loses its preload or a belt loosens up. Suddenly, the resonances shift in frequency and the notch filters no longer do the job. The autotuning process has to be repeated.
Ideally, the operator notices the problem and takes action but that may not happen immediately. In the interim, machine throughput and performance start to suffer; first only incrementally but more over time.
The latest crop of autotuning drives is designed to take care of the problem so that the operator doesn’t have to. These drives monitor the performance of the axis in real time. When they detect changes, they rerun the autotuning routine to optimize operations as much as possible. In the best case scenario, by autotuning components they may return performance back to baseline. Long-term, this is not a solution for failing components, however, this is why they also notify maintenance of the change. That way, the root problem can be diagnosed while it is still minor, allowing maintenance to fix it during a convenient window that does not impact operations. Look for broader deployment of these types of autotuning drives in the coming years.
Big Data in the Factory
All of the sensors we’ve been discussing potentially generate a dense data stream around the clock. Obviously, this stream contains valuable information. The question is how to sift through it to find the valuable nuggets and how to do it quickly enough that the results actually deliver benefit. The answer is to use big data techniques.
Although the dialogue around big data tends to focus on consumer-centric applications like healthcare and retail, manufacturing promises to become one of the biggest beneficiaries of the technology. The companies who gain a competitive advantage will be the ones who acquire the technologies and skill sets to leverage the information effectively.
It starts with having a plan. Collecting data just for its own sake not only fails to deliver benefit, it actually creates a problem in terms of the cost of acquiring and managing storage assets and computing equipment. Organizations need to first define the problem they wish to solve, then determine the data that requires. Someone needs to take ownership and responsibility for maintaining it.
Big data also requires serious computing resources, along with IT staff with experience in Linux and specialized applications like Hadoop and Spark. There is currently intense competition for these individuals but expect that to ease over the next four or five years as the current crop of students begins to graduate and complete internships.
The Industrial Internet of Things
All of this discussion brings us to the IIoT and its fabled $280 billion of connected devices. This doesn’t just involve sensors but smart motion components, communications nodes, environmental controls and more. For all the discussion of controlling home thermostats via smart phones, the IIoT promises to dwarf that with 5.4 billion industrial IoT devices shipped annually, on a 48-billion-unit total installed base. The result of all of those devices? Greater transparency up and down the food chain:
- Less downtime
- Shorter lead times
- Reduced inventory levels
- Easier customization
- Better efficiency
Let’s take a closer look at an example. Oil prices may have plummeted over the past year but manufacturers continue to focus on cutting energy usage as a means to both increase profits and decrease environmental impact. New approaches to energy monitoring provide a method to do just that. Energy bills for businesses have two components: usage charges and peak-demand charges. Pushing more power out to a facility with higher peak demand requires more robust infrastructure, from power lines to transformers. Utilities recover some of the cost by scaling the overall usage charges of a specific facility by the peak demand over the previous month or quarter. Peak demand charges represent as much as 30% of energy costs. Lower the peak demand and you lower your energy bills (see Table 1).
This is where the sensors discussed earlier come into play. By monitoring operations, management can determine the best strategy for lowering overall energy costs. Rather than running the line at maximum speed during the middle of the day, for example, they can push those operations to the night shift. Alternatively, they may be able to produce the same amount of goods by operating longer at lower speed. Of course, this involves a trade-off-- paying staff for longer shifts or nighttime and weekend hours may counterbalance the savings. Still, as the table shows, even a small change can have big impact.
In order to make these kinds of trade-offs, manufacturing organizations need to know exactly what’s going on in their operations. This is where the kind of sensors and analytic capabilities we’ve been discussing come into play. It’s possible to build these frameworks from scratch but then there’s increasingly offer the integrated equipment designed to simplify matters. Data loggers extract data in real time, interfaces to manufacturing execution systems simplify putting information in the hands of decision-makers.
The technologies discussed above lead us to perhaps the single most important trend over the next several years – the shift toward the intelligent factory.
The combination of sensors, the broad connectivity of the Internet of Things and big data functionality will increasingly enable organizations to mine their data for actionable insights. This level of transparency into operations will enable the entire organization, from machine operator to upper management, to make decisions that will streamline performance and enhance profitability.
There was a time that factory monitoring was both limited and laborious. Was a piece of equipment running? How many parts was it turning out per minute? Were the various elements operating within bounds? Data had to be gathered and manually and analyzed off-line. The information was somewhat useful but often outdated by the time it was available.
Enter the intelligent factory. Depending on the implementation, information can be available in minutes or even in real time. Now, management can review information like operational equipment effectiveness, throughput and changeover time. Maintenance and system integrators alike can have parameters like current draw and load curve delivered to their mobile phones.
That’s the upside. The downside is that it can add cost, not just for capital equipment but for both integration and programming. The learning curve can be fairly steep. In response, a number of vendors have launched intelligent factory initiatives with dedicated product suites aimed at helping their customers derive value as quickly as possible. These purpose-built factory-monitoring systems are designed to work with the equipment to gather information in real time, export it and deliver it in a highly consumable form.
Instead of each facility having its own separate factory monitoring system, a single solution is replicated and available across the entire organization. Systems can be configured to display information to machine operators so that they can easily see anomalous behavior and take action. Data loggers and interface equipment automatically extract machine and operational data and relay it to management systems, whether on computer or mobile platforms. Management has access to analytic information across all facilities, allowing them to assess and compare performance from facility to facility, line to line, machine to machine and even operator to operator. If a specific facility or shift outperforms the rest of the organization, management can discover that immediately. Even better, they can now determine how and why that happens and export the techniques to the rest of the company.
“Manufacturing execution systems and the intelligence provided lead to better knowledge, insight and actionable decisions that improve workflow, quality, and productivity,” says Janice Abel, principal consultant at ARC Advisory Group (Dedham, Massachusetts). “The solutions are becoming more intuitive, situational and location aware. They offer the ability to compare best practices and functions across sites, product types, or make changes in real-time.”[Reference 3]
These are essential trends that will be covered on the Motion Control & Motor Association over the coming year. Watch for detailed articles on the status and developments these areas. For more information on the trends presented here, download the webinar here.
1. Total Available Market for the Industrial Internet of Things Global Market Research Study, ARC Advisory Group, May 2015.
2. Sensors empower the new-age industrial revolution, Mergers Alliance, June 2015.
3. Industrial IoT and Analytics Drive Growth of Manufacturing Execution Systems for Discrete Industries, ARC Advisory Group, March 2015.