Making production data work for you

By Ron Iannacone, President, FIN, LLC September 21, 2005

Each year, manufacturers lose nearly $20 billion — or almost 5% of total production in North America — due to unscheduled downtime. Despite the increasing sales of data collection systems, this loss suggests that manufacturers still don’t know what is causing line problems and equipment failure.

Over the last four to five years, manufacturers have invested heavily in sophisticated technology such as HMI and SCADA packages to help them better visualize and control the manufacturing process. However, these systems are designed to be an operator interface to your control system, and often have limited data collection and storage capabilities. Additional components and engineering may be needed to glean the data from these systems that manufacturing personnel need.

And what if you don’t have a control system, or are using legacy equipment that doesn’t have a PLC? Or what if your plant has no engineering staff, and can’t afford SCADA systems or the personnel required to set up and run them? Both of these factors prevent plant personnel from getting the data they need to make real productivity and product quality improvements.

Intelligence-driven decision-making

Many manufacturers don’t realize what data collection really is, and how collected data becomes usable information. Data collection has become a catch-all phrase, covering everything from logging machine temperatures and pressures by hand, to sophisticated systems that display and record real-time data from hundreds of data points, to control systems equipped with data collection capabilities that store data in multiple, disparate factory and enterprise systems. What’s usually missing is how to get that data in the right context to the right people at the right time to provide support for better decision-making — true factory intelligence.

An emerging class of manufacturing system called “factory intelligence” could bridge the gap. These systems are designed to give plant personnel quick access to critical process information. Factory intelligence systems can replace manual data collection systems, or work along with existing control systems, data historians, HMIs, EAM systems, MES, PLCs and I/O modules to collect and store data from both legacy and new production equipment. Factory intelligence systems include the necessary hardware interface for each piece of equipment if it is not already so equipped.

Factory intelligence systems continuously monitor and record all critical machine functions including motor current, motor speed, faults, temperature, speed and other factors like pressure plant air and/or hydraulic pressure, electrical, and water usage. Plant management can set ideal parameters so that alarms trigger when changes occur outside that range, and the events are logged and stored by date and time for future trending and analysis.

At any time, manufacturing personnel can access this stored information to troubleshoot line issues. Managers can analyze and trend the information to make smarter purchasing decisions or process changes to improve quality. For example, suppose a large percentage of a specific part is coming back defective. Managers could use the system to bring up the operating parameters of the specific machines used to produce that part, and see the conditions present at the time the part was made. They could use this information to determine the precise cause of the defect and the action necessary to resolve the issue.

So collecting data is important, but it is not enough. Manufacturers must have the means to turn that information into intelligence they can use for on-the-job decisions. They must have a way to continuously collect and record that data. They must have a user friendly interface geared for use by everyday plant and maintenance management and personnel. Most importantly, they must have a way to store that data for the long-term. Factory intelligence software is an affordable way for small and mid-sized plants with 200—5,000 data points to achieve all this.

Making legacy equipment smarter

One of the biggest challenges with data collection is capturing ALL critical data from ALL equipment on the plant floor. Most small to mid-sized plants have legacy equipment (here defined as having no PLC). Legacy equipment was expensive when purchased, is still reliable, functions at or above satisfactory levels and is now paid off. Manufacturers know they want to keep this equipment, but they often have no means to collect process information from it. And that means they are receiving an incomplete picture of what is taking place within their operation. Factory intelligence systems can collect data from legacy production equipment as well as from discrete systems like electronic scales.

Many manufacturers have implemented ad hoc manual data collection systems. Although better than nothing, these systems are tedious, unreliable and expensive. For example, a manufacturer of plastic pellets requires operators to write down process parameters every hour during a production run. Process data is then entered into spreadsheets by hand. Running three shifts per day on five lines would consume approximately 14 man-hours per day; the cost would be approximately $52,000 per year. And the accuracy of the data could possibly be questioned. Also, in the event of an incident, this data is not easy to view and trend. Alternatively, a factory intelligence system can automatically produce run records every shift on an internal web page. Accuracy is assured and the data is stored securely and indefinitely.

The need for continuous data collection

Continuous monitoring means that you don’t miss a single event so you can access this data when equipment failure leads to unscheduled downtime. Despite the best efforts at preventive maintenance, equipment breaks downs as a result of mechanical, electronic, software application, operating system and human error. When equipment fails, personnel can use the factory intelligence system to pinpoint the operating parameters and machine conditions present at the time of the failure, as well as minutes or even hours prior to the event. You can see what went wrong immediately, preventing maintenance staff from wasting precious production time trying to fix the problem and get the line running again.

Continuous monitoring of equipment operating conditions such as speed, temperature and pressure can also help managers anticipate, and therefore prevent machinery failure. Personnel can use the system to get a complete picture of operating conditions, and to flag events as soon as they fall outside of the ideal parameters; they can take corrective action before the event causes a shutdown.

For example, the incident in Figure 1 occurred on April 7, 2003. The machine operator called maintenance, reporting occasional high temperature alarms on the calendar rolls that seemed to happen only when they were adjusting the heated calendar roll gaps. The maintenance manager went to his computer terminal, quickly trended the temperature of the roll versus the operator hitting the gap open or closed. The trend shown in Figure 1 leaves no doubt that the operator was correct. This allowed maintenance to determine in a matter of minutes that the thermocouple had a signal problem that was associated with the gap motor turning on.

Most operators assume that sophisticated HMI and SCADA systems continuously monitor all data points. But these systems are designed for control and real-time visualization, not continuous documentation of machine parameters. As a result, there may be gaps as long as 30 to 45 seconds between data recorded from a given point. It may not seem like much, but a lot can happen during a 30-second interval. For example, using factory intelligence software, an HMI/SCADA system at Tarkett Flooring in Vails Gate, NY provided no clue to what was causing major product defects because it was capturing data only at 30-second intervals. Using the factory intelligence system to view and trend the same events, Tarkett was able to clearly see that multiple, one-second speed variations were causing the defects.

A system designed for the common man

Data historians collect data, but are not an option for all manufacturers — particularly small-to-medium sized operations. The main reason is that data historians can interface with only SCADA, HMI and/or PLC-based systems. And many manufacturers cannot afford these systems or the engineering staff required to install and operate them. Even if the operation does have a data historian, the information may not be very accessible to plant personnel. It can be very cumbersome to extract and store data in a way that it is readily accessible when needed in the future. Factory intelligence systems are specifically designed to make this information readily accessible to plant personnel.

With a factory intelligence system, you can walk into the plant, go up to a computer, pick line speed, motor current, pressures, tensions or virtually any other parameter and see when a line went down. You can watch the speed go from X to zero, see what your process parameters were at that moment — even right before or for the past day. You can trend these points along with an E-Stop, limitswitch or fault contact and see if they changed state at the exact moment the line went down.

When Chris Kasmer, Operations Manager at LTL Color Compounders in Morrisville, PA wanted to check trend data for one of his new or legacy extrusion lines, he had to physically go out onto the plant floor and displace the machine operator at each line control console. He knew the information was somewhere in his SCADA system, if he was willing to go through the cumbersome process to view and trend it. Plus he had to work with the data in a noisy environment that was not conducive to data analysis and planning.

Kasmer now monitors real-time production data from his office PC using factory intelligence software. He can view the data for the entire operation, or for each individual line. He can pull up historical data, trend it, organize it, view it, document it and store it according to his needs. He can access the system via the Internet from anywhere in the world. During plant emergencies, such as unanticipated shutdowns, he can troubleshoot and offer possible solutions — even remotely, when he is offsite.

Long-term storage capacity

Historical data is vital and can serve many purposes. But its value is limited if it cannot be retained permanently. As mentioned earlier, most systems are only designed to act as an operator interface to your control system and to monitor data in real time, which means they have limited data storage capabilities. They don’t store the historical data needed to troubleshoot faults, machine failures and process issues. And that means that in the event of a line failure, information about past events cannot be retrieved.

Factory intelligence systems ensure that plant personnel can access critical operating data that is weeks, months or even years old. They can call up data on process parameters such as raw material usage, heat and pressure, as well as machine faults, utility usage and even environmental factors that can affect production and product quality. They can pinpoint the process parameters that caused a defect to appear in a customer’s shipment several months prior.

For LTL Color Compounders, maintaining the specific parameters of its production equipment is critical to maintaining quality control and customer satisfaction. “Many of our customers have products with extremely narrow tolerances for color deviation,” said Chris Kasmer, operations manager at LTL. “That’s why we need to record parameters for each batch run of product. This gives us the information we need to replicate colors when customers reorder.”

Proactive maintenance

Factory intelligence systems also help manufacturers practice proactive maintenance. Figure 2 shows a 24-hour trend of four bearing temperatures on a mill. Maintenance monitors this trend in real time and ensures that the bearings are greased quickly. The green and yellow lines are bearings that have caused lubrication problems. Using a factory intelligence system, maintenance could immediately see the temperature increasing above 350 F on one of the bearings. As a result, they knew that the bearing needed to be lubricated, thus averting unscheduled downtime. The trend shows the resulting decrease in bearing temperature. Maintenance can also set alarm limits for the bearing temperatures to prevent future failures and schedule maintenance as needed instead of during planned shutdowns.

Another approach is to monitor key control loops to see how stable they are. Control loops may appear to be stable, but are actually oscillating, which can cause premature failure of components. Valves can be trended to look at the time between opening and closing to see if they are sticking. Motor current, bearing temperatures, speed irregularities, etc. can be constantly monitored for wear and degradation of performance. Manufacturers can even monitor air or power abnormalities. The bottom line is that you can schedule shutdowns based on actual equipment status, and focus maintenance only on the equipment actually requiring service.

In this era of lean manufacturing, many sophisticated data collection systems are priced beyond reach for small and mid-size plants. The costs of large-scale custom configuration can reach upwards of $500,000. Plus, once installed, these systems often require the expertise of an on-site systems engineer to configure and operate. Factory intelligence systems are pre-configured by systems engineers based on on-site surveys to work out of the box. They require minimal training, and are designed for use by plant personnel.

For more information, visit www.factoryintelligence.com or email riannacone@factoryintelligence.com