Smart Machinery: Critical to Running Your Business

By on September 7, 2018 in PROCESSES & FABRICATING

Many machinery production lines continue to use aged manufacturing processes with little integration of modern data analytics and decision-support tools.

The history of materials is filled with transformational advancements catalyzed by increasingly lighter materials that have made it possible to achieve more functionality with less. Recent fuel-efficiency improvements in the automotive and aerospace industries and material cost savings in industries like RV and marine are examples of this trend.

As newer, lightweight materials become more common, new manufacturing processes have to be developed. Often industries rely on long-established traditions of product development and manufacturing that may have to be adapted for these new materials. As changes are made, validating the innovation is an improvement over conventional processes is critical to justifying the new systems. This need has opened the door for comprehensive data reporting for process improvement, decision support and validation. The data obtained provides deeper insights and allows for continuous process optimization.


In the United States today, many machinery production lines have been in service for decades. Often, they continue to use aged manufacturing processes with little integration of modern data analytics and decision-support tools. While the processes were most likely very innovative at the time, and may have even seen upgrades over the years, most were not designed to give deep insights into the process midstream. This is not a design flaw, but a reflection of an age where processing power and sensors were far costlier than they are today.

Understanding what is going on with individual machines at any given point is traditionally an expensive endeavor that oftentimes requires data to be recorded manually. Manual data is slow to record, not in real time and expensive to produce—meaning very little process improvements can be made in a timely manner. One solution is to implement a custom system designed to automatically record preselected data points. This approach limits flexibility for adding new machinery, offers only a retrospective review of the machining process and is often cost prohibitive.

Looking at one specific piece of machinery in context of a factory gives a better understanding of how data fits into the framework of the whole production process. The CNC Machining Center is a practical example of a versatile machine that is used in many manufacturing situations. With the CNC at the center, there are numerous steps before and after the actual machining stage. Initial stages may include: preparing materials to be machined, forming a mold, laminating a sandwich of multiple materials or other such steps that generate the part that will be machined. The CNC operations are then carried out on the machining center, producing a part or series of parts.

Once the CNC machining is complete, assembly and finishing take place. All these operations may occur either in-house, for more vertically integrated factories, or be outsourced for more focused manufacturing facilities (orsometimes a combination of both). Each step of the process creates data points that can be relevant to optimizing the whole system.

Fig. 1 – Osync UI

To understand the entire process, a standard method of reporting information should be used. This is where the concept of Industry 4.0 comes into focus. Industry 4.0 is considered to be the fourth Industrial Revolution where cyber-physical systems coordinate the passage of information among each other to improve the overall process. While the actual format of the reporting will be tweaked to the process that is being observed, a common communications protocol is needed to let various software collect and display the data adequately.

Two widely used industry standard protocols for machinery are MTConnect and OPC-UA. These provide a framework for reporting information in a standard format that many applications can understand. These frameworks are also extensible for custom-reporting needs. As an example, Osync Machine Analytics®, produced by C.R. Onsrud, Inc., has extended the standard MTConnect protocols to provide additional information that is helpful for businesses to understand the processes related to their CNC machinery. (See Figure 1.)

Partners of C.R. Onsrud are serving up information via the MTConnect protocols so that surrounding devices like vacuum pumps, aggregate tools and other accessories have their status monitored in context of the overall process. Additionally, information can be passed from upstream processes such as CAD/CAM or PLM software that may be relevant to the operation. By connecting these accessories and upstream systems, a clear picture of the entire operation can be assessed. The information may even be integrated with an ERP software to fully connect the process to the larger manufacturing picture.

Closed platforms are designed to handle a limited set of assets and use cases, which means it can be difficult and costly to expand.

Fig. 3 – Visualizing your data in different formats can be key to gaining insights into your equipment’s operation.



When monitoring equipment, it is imperative to use an industry standard communications protocol, ensuring maximum interoperability over time. If one vendor stops offering the solutions needed, there will be another vendor to move to, without having to re-create the entire system and migrate all data to a new platform. Additionally, closed platforms are designed to handle a limited set of assets and use cases, which means it can be difficult and costly to expand. Open platforms have the best combination of products for each use case while tying it all together with a common communication language.

When monitoring processes, there are many aspects to consider. Real-time data is useful for providing continuous awareness of ongoing processes. Being able to see transient issues or stoppages, as they happen, is key to improvements in daily efficiency. To catch these issues as they occur, a real-time notification system is useful. These alerts are designed to get the attention of a supervisor who may not be actively monitoring the process. Real-time data is extremely valuable, but on its own, only provides a snapshot in time, at the moment it is observed, and is often less useful for evaluating the impact of less common events or small system changes.

Recording real-time data so that there is a historical log of information provides deeper insights. The historical information can be correlated to help predict future jobs outcomes. Additionally, historical data provides a better picture of the frequency of a problem.

Summary information, on the other hand, is useful for evaluating trends and the overall impact of changes on outcomes. It is often difficult to appreciate the impact of a change until summary data is available for full evaluation. Summary data is one of the best indicators of where to focus improvement efforts. A comprehensive solution including both real-time and summary data, along with a historical reference, is key to having the business data needed to run a smart factory.

However, it is not all about the process data. A big part of productivity is about keeping the machine running in its optimal state. There are several levels of machinery maintenance. The most common is corrective maintenance. This is where a machine is run until it fails or has an error. While this may ensure that it is not taken out of service unnecessarily, it may also mean downtime at an inconvenient or critical point. This method often requires that there are many readily available spare parts as it is difficult to predict when something will fail, resulting in a significant downtime.

The next level is preventative maintenance. This is centered around doing routine checks, based on an average schedule, to determine if anything is abnormal. This avoids unpredictable catastrophic failures. However, it comes at the expense of having to schedule routine checks—which leads to underutilization of the machinery and additional labor costs for technicians to check the equipment.

The ideal level is predictive maintenance. Predictive maintenance uses sensor data combined with statistical analysis and trend monitoring to alert when a part is predicted to fail. This allows for a maintenance window to be scheduled, parts ordered and delivered, all with the least impact to the production process. With predictive maintenance, the part is replaced before it fails. There may be a little bit of useful life left in the part. However, the downtime costs far outweigh the minimal useful life left.

This process has not been historically used as it is very data intensive. A great deal of data must be collected to build the statistical models and correlate what failure modes are relevant. Additionally, a lot of in-process data must be continuously recorded in order to compare the trend to the model and determine that a failure is imminent.

Keep in mind, everything cannot be predicted. For practical purposes, an efficient factory requires a combination of quality construction, predictive maintenance, basic preventative checks and a plan for corrective maintenance covering unforeseen events. (See Figure 3.)


Data has become one of the key defining factors in many businesses. It gives an understanding of where bottlenecks are so efforts can be focused on optimizing the items that could have the most impact on the business. If the business is automatically generating the data through machine monitoring and the daily tasks that are happening anyway, it is the most efficient way to gather insights into processes.

An efficient factory requires a combination of quality construction, predictive maintenance, basic preventative checks and a plan for corrective maintenance covering unforeseen events.

One great way to use data to help the business is by looking at historical information. Future job costs and projected revenues can be extrapolated by comparing to past data. Jobs with similar characteristics are correlated based on their run times, job setup or changeover times, as well as key run statistics, to help formulate an estimate for future work. By using information to understand the real-world margin on jobs, future jobs will be bid more accurately.

Job-cost analysis is further enhanced by factoring in the actual machine monthly payment (if financed), operator hourly pay, hourly overhead, consumable costs, etc. This gives a true cost model per job instead of a general overhead model. This more accurate reflection of the expenditures associated with a job permits modeling adjustments to the process to see what sort of gains would be achieved if various aspects were modified. Some items that are commonly refined are shift length, additional shifts, tooling, load and unload times, and machine utilization, among other factors.

The metric of overall equipment effectiveness (OEE) can be used for optimization between similar types of machines. OEE is calculated by measuring the availability multiplied by the performance multiplied by the quantity. Availability is defined as run time divided by planned run time. Performance is ideal cycle time multiplied by total count and then divided by run time. Quality is defined as good count divided by total count. OEE is used to determine where one operator may be slowing down the process compared to another operator or where one machine may be producing more nonconforming parts than another. Based on the compared metrics, the process can be adjusted appropriately.

When looking to optimize a factory process, it is best to use a software that visualizes each step and provides information in a clear and easy-to-read manner. This data should support optimal utility of assets while providing actionable business intelligence, not just mounds of data to comb through. The more advanced software, like Osync Machine Analytics®, will even provide specific adjustment recommendations to support a better result. Knowing the right lever to pull in order to adjust a process gives a business a key advantage in an increasing competitive landscape.


Consider machinery’s ability to report analytical data when purchasing.
Investigate if the machinery provider uses industry standard protocols and an open platform to allow for expansion and flexibility.
Evaluate the software solution for inclusion of real-time, historical and aggregate data for both job and diagnostic information.
Ensure data from the solution is reported in an easy-to-use, actionable format to support timely understanding and decision-making, rather than a data dump.


By Jeff Onsrud, Director of Sales & Business Development, C.R. Onsrud, Inc.

Jeff Onsrud is director of sales and business development for C.R. Onsrud. He has been in the machining industry for more than a decade in various roles from technician to machinery designer to sales engineer. With a degree in electrical engineering, he has always kept a pulse on the IoT landscape. With sensor cost coming down and machines becoming more connected, Mr. Onsrud makes sure C.R. Onsrud stays on the forefront of these technologies by leading the development of products like Osync Machine Analytics in order to provide more business value to their CNC machinery customers.