Predictive and preventive maintenance tools spot potential disturbances in critical production equipment before they turn into major headaches, allowing food and beverage plants to repair, rebuild or replace an asset on a schedule rather than facing an unplanned line shutdown. Consequently, processors are moving from calendar-based checks and reports to condition monitoring of their assets in real time.
“Food and beverage plants are realizing the failure of a critical asset in a production line can add thousands to hundreds of thousands of dollars in lost productivity, even if the disturbance results in just a minute of downtime,” says Katie Moore, global industry manager for food and beverage at GE Intelligent Platforms.
The run-to-failure approach, at least for critical equipment, has given way to collecting data that can warn of a problem days, weeks or months in advance of it happening. “Capturing and organizing the available data from assets on the production line and in the plant, and then giving the right data to the right people at the right time help the technician make better decisions about the asset,” adds Moore.
But sometimes, run-to-failure thinking is not necessarily a bad strategy. Some assets of a certain size are continuing to be intentionally run to failure and then replaced, says Terrance Fisher, global strategic account manager at SKF USA. “Choosing to let the asset run to failure indicates a shift in thinking; plant personnel are aware and make a proactive decision depending on the criticality and the economics of the asset.” One major food maker’s embrace of SKF’s PdM tools is so strong, it now has trouble getting its plant maintenance teams not to schedule the repair of small electric motors that are plentiful but not critical in the plant.
Newer predictive maintenance (PdM) and preventive (PM) tools trace production problems back to previously hidden sources, such as having a motor no longer suited to a particular application because a different product is being run. The latest predictive and preventive tools bring operator and maintenance technician input into the system where their experience plays a larger role in managing equipment maintenance. At the same time, lower-cost and more affordable yet robust PdM/PM versions are becoming available, targeting mid-sized firms looking to reap the benefits of the technologies. Other developments bring new types of equipment and materials under the scope of PdM tools.
While gear boxes, pumps, motors, mixers and centrifuges are the most common critical assets in food or beverage production, lubrication, piping, valves, evaporators, compressors, condensers and conveying belts also are being pulled into the scope of PdM.
Predictive maintenance involves instrumentation or diagnostic tools like an infrared (thermal) camera, vibration or oil analysis or ultrasonic device to capture a reading and get a condition check, says Fisher. “Whereas preventive maintenance could be as simple as changing the oil in a gear box on a weekly or monthly basis, it is done to prevent a failure from happening, but it really doesn’t have a specific diagnostic nor does it rely on preventive technology to tell someone when to do it.”
The standard measure of a PdM program’s success is defined by completion—whether it was done or not done, says James Freaner, director of sales at Advanced Technology Service (ATS). But some people say the definition should be narrower, where the food plant asks if the PdM resulted in a reduction of failures to a particular asset and, if not, how the plant will adjust the PdM to address the root cause. “What a food plant wants to know is whether its PdM strategy is reducing equipment failures, overall downtime or the occurrence of reactive failures,” says Freaner.
The challenge is to drive an enterprise-level initiative around maintenance and reliability, explains Freaner. “At ATS, we’ve been able to document seven-figure savings across an enterprise with the implementation of a proactive maintenance solution. Depending on the number of plants, successful replication of a proactive maintenance solution across the enterprise becomes a major contributor to the bottom line of the business.” The question now is what technologies, tools and best practices help food and beverage firms continue to drive out reactive break/fix and become more proactive as an organization to reduce unplanned downtimes.
“Predictive maintenance programs, if set up correctly, should be able to determine catastrophic failure of an asset,” says Chris Hanssen, director of customer service at JBT FoodTech. “If we go back and analyze data after such an event, we can see if an operator is simply resetting the alarm, restarting the equipment and running again without finding the root cause of the problem. Not heeding an alarm that leads to catastrophic failure of an asset may not seem as serious as a recall, but it could be the second-biggest profit-reducing event on the list,” says Hanssen.
Typically, when an unplanned breakdown occurs, the urgency is placed on restarting production, so maintenance personnel have no time to discover the root cause of the problem. A technician simply finds the least time-consuming solution so production can quickly resume, says John Kravontka, president of Fuss & O’Neill Manufacturing Solutions LLC, an environmental and civil engineering firm focused on the reliability and safety of manufacturing equipment.
Get real time
GE’s Moore says real-time delivery of production line information is essential for a predictive or preventive program to deliver maximum results. “Several different types of data can be captured including time series data, where a check of the status of the equipment or a measurement taken from a piece of equipment is associated with a time stamp. Equipment performance data is captured periodically also, such as when a HACCP plan at the plants needs it, or an FDA compliance guideline requires it, or the EPA or corporate management needs it,” says Moore.
GE Intelligent Platforms’ Proficy suite of integrated data management and data analytics software captures, archives, contextualizes and distributes volumes of real-time information and allows the user to determine the root cause of process upsets. The Proficy suite includes a newly developed Real-time Operational Intelligence (RtOI) paradigm that empowers businesses to transform vast amounts of operational data into actionable information that can be accessed on mobile electronic devices. Moore says RtOI can capture the tribal knowledge of the operator on the plant floor in addition to the performance data on an asset.
An operator walking by a pump can add a note to the maintenance system if he or she hears an unusual noise, signaling a potential out-of-tolerance condition. Because the information can be captured and tied back to the equipment, the maintenance manager sees the note when data on the asset is pulled, says Moore. If an operator walks up to a piece of equipment with a tablet in hand, the maintenance system knows who the operator is, his or her role and position, and the individual’s location, training and level of access to the asset. The system prioritizes the actionable events—alarm conditions personnel can resolve—pertaining to the specific asset, sending the information to a tablet or smartphone automatically.
The suite also includes Proficy SmartSignal predictive analytics and diagnostics software that can automatically and continuously identify what asset is going to fail and other impending equipment problems to avoid unexpected shutdowns and catastrophic failure.
The software sends a brief summary notification to designated individuals, identifies the priority level of the impending failure and determines the apparent cause. The software also helps users organize maintenance projects and leverages existing instrumentation, tools and IT infrastructure in the plant.
In addition, the system compares historical production data to real-time data to estimate the severity level and importance of the maintenance required. “If you can look at historical data and what is happening in real time, like temperature and humidity, and you see one of those two drifting, you can alarm/alert before it reaches a point of no return and correct it. This proactively ensures your food safety,” says Moore. Users can tie a wide range of data back to the processing parameters happening at the time and see what effect any step change in value will have on the end product. “Operational excellence begins with understanding your data. Tying that information together starts to build a greater holistic view of what is going on in the process from maintenance to manufacturing,” according to Moore.
PM for mid-sized plants
SKF offers a maintenance management service to small- to mid-scale manufacturers that allows them to reap the benefits of predictive maintenance without investing in the related equipment or specialized data analysis training that a PdM program requires.
“Start-up costs for a PdM often exceed $100,000 per year, so we have devised an approach for smaller firms. Smaller plants are understandably focused on production costs and, therefore, fail to consider the cumulative, bottom-line impact of lost production attributable to a breakdown,” says Fisher.
SKF’s Machine Health Reporting Program, commercialized in 2010 and part of SKF’s 360° Solution, is a partnership between the company and client whereby SKF trains maintenance personnel how to use its handheld data collection technologies to capture machine data while maintenance performs its normal duties.
Maintenance personnel at the plant collect the data, which is then sent to a certified SKF Reliability engineer for analysis. The SKF engineer identifies problems and suggests actions to avoid unplanned downtime. SKF issues an alert if an urgent condition relating to a piece of monitored equipment arises. SKF publishes a monthly report on the health of assets and posts it on a private web page, says Fisher. Users of the machine monitoring program pay a monthly subscription fee based on the number of machines being monitored. (It’s a tiered system where 100, 250 or 500 pieces are monitored.) Fisher says a typical food plant monitoring 250 pieces of equipment spends about $25,000 annually for the service.
SKF also rolled out a new Documented Solutions Program (DSP) that shows how a PM or PdM system user can extend its meantime between failure or calculate ROI for any piece of predictive or preventive maintenance that gets done. SKF inputs customer data including hourly rate and the cost of downtime, components and the replacement of components that fail.
“Maintenance can use the results for the corporate purchasing or corporate engineering organization to justify an expense on the maintenance side and to show how PdM or PM can increase profitability for the company. Hypothetically, a manufacturer operating at 16 hours per day, five days a week could realize a potential productivity gain of $261,000 annually if it experienced a meantime between repair of 48 months, monitored 250 machines, had an average expense of $1,000 per repair and cost per hour of stopped production amounted to $1,000. “Unplanned downtime can include a range of fixed, variable and industry-specific costs that drag down productivity and profit,” says Fisher.
Full asset management use
Asset management systems at many plants are only partially deployed, says ATS’s Freaner, and are not doing enough to help offset repetitive failures. This may be due to the practice adopted by food plants to operate their preventive maintenance systems on a calendar-driven basis, where checks are run monthly, quarterly or semi-annually. “The PM should be modified to be driven by machine run hours and machine run time, which is a better indicator in that it provides a more immediate, realistic view of machinery conditions,” says Freaner.
ATS developed eFactory Pro, a web-based CMMS system powered by SAP, which handles all aspects of factory maintenance including materials, work order administration, metrics and planning/scheduling. Leveraging the ERP expertise of SAP enhances the flexibility and user friendliness in deploying the CMMS and in the predictive and preventive information and data pulled out of it.
The ability to correlate maintenance metrics to production metrics helps maintenance contribute to the plant’s productivity and efficiency. Standard maintenance metrics, including downtime, availability, meantime between failure and meantime to repair must correlate to machinery uptime. Then the plant has to measure whether moving the needle in the maintenance arena is having an overall effect on plant productivity, says Freaner.
eFactoryPro features a suite of advanced standard predictive tools for asset monitoring by thermography, vibration monitoring, oil analysis or ultrasound, and adds a new capability that allows maintenance staff to generate a work order when they detect a potential problem. The work order is generated by an engaged technician who is not just looking to repair the problem at hand, but to identify any other potential problems and address a root cause in a way that prevents downtime.
For example, if a technician executing a PM task on a piece of equipment notices a hydraulic line is wet, and he or she doesn’t have time to address the problem right then, a work order to begin the process of scheduling evaluation and repair can be started. Then it is added to the back log for that particular asset, and the plant can assign a technician, coordinate materials and block time on the machine as opposed to having a failure and unplanned downtime.
Manage your micro stops
Maintaining and keeping the uptime of a manufacturing execution system (MES) or recipe management or batch data system at a high level is more difficult to do, since there is so much activity on the factory floor from a database and PC/IT perspective, says Matt Ruth, vice president at Avanceon, a total systems integrator. But PM and PdM can help. The growing demand for a PM/PdM strategy manifests itself in numerous support contracts.
Predictive and preventive maintenance is a service Avanceon provides where it gathers information from the historians and logs of an existing system in a facility. Then, predictive and preventive maintenance is run off of the analysis of the data as part of its support services organization.
Avanceon identified new areas in production processes using advances in its predictive technologies that enhance overall equipment effectiveness (OEE). Standard OEE and performance management systems track downtime events when they occur and, along with predictive programs, can be used to determine the top reasons for downtime.
However, OEE often overlooks events that Avanceon calls micro stops, productivity slowdowns rather than catastrophic failures that occur when the process is in a steady state. “They are the hiccups and glitches that keep you from running at peak performance,” says Ruth. While they do not trigger a shutdown and do not have a long duration, micro stops represent a large amount of inefficiency when summed together, says Ruth.
The micro stop concept upped productivity at a food supplier when it identified that a particular type of material from a certain vendor was causing mini jams on its production line, resulting in lower product output. The material did not result in downtime on the filler, but staff had to be there to clear the jams. The micro stop concept provided the clues, allowing the food maker to contact the supplier and fix the issue. “It’s more predictive supplier management and less predictive maintenance, but it drives down the chance that operations run at a reduced production rate,” says Ruth.
Purge generic PdMs
A number of food plants lose out on potential productivity gains by incorporating generic PdMs into their maintenance programs and not trying to optimize them. Even worse, some plants cut and paste a PM from one machine to another, says Fuss & O’Neill’s Kravontka.
When this approach is in play, nearly half the time spent on preventive maintenance is wasted working on equipment that did not need to be checked but was looked at because the PdM was there. Approximately 10-15 percent of unnecessarily scheduled PdMs cause damage to the equipment being checked, likely due to an error in installing a replacement part. A piece of equipment that performs worse after a maintenance check will cause a production supervisor to hesitate before agreeing to schedule maintenance the next time, says Kravontka.
Fuss & O’Neill helps clients in food and beverage develop a PdM strategy focused on their critical equipment, mating those production devices with the right predictive tool.
The company evaluates the client’s maintenance organization against a set of standards. Fuss & O’Neill can determine what a company does well and what it does not do so well. It develops a strategy specific to each company and recommends certain technologies and tools.
“Users have to understand the PdM is designed to find problems, so it has no value if it does not find them. Corrective action work orders are one way preventive programs find potential failure while the problem is still in its infancy and a less complex fix. If you do not generate corrective action work orders to repair those problems, that is a missed opportunity,” says Kravontka.
Kravontka says condition-based preventive maintenance is growing because it can deliver better information compared to time-based approaches. “If a food plant changes the oil on a critical piece of equipment once a year and has hundreds of pieces of this equipment, a switch to condition-based monitoring using oil analysis can save a lot of money. With this type of monitoring, the equipment owner takes an oil sample and lets it direct what to do next. The decision to change the oil is based on its condition; if the oil is still good, no maintenance on the equipment is required.”
Preventive maintenance practices and programs have had an unintended effect of creating maintenance schedules on assets that may or may not require maintenance, says Steve Hawkins, Eastern region manager-refrigeration services at Stellar. “It is important to look at a PM and determine if it is being done for the sake of doing it or because the asset truly needs it, because downtime for maintenance has to be scheduled. Touching an asset that is working well and does not need maintenance introduces the chance for error and failure.” This practice has led to more PdM and condition-based monitoring to ensure the maintenance strategy is effective.
Users of predictive maintenance programs may be able to postpone preventive maintenance on a particular piece of equipment. “If a compressor is scheduled for rebuild every 30,000 hours, predictive technologies such as a vibration and oil analysis can verify whether there is a change in the signature or health of the equipment that would signal an oil change and cleaning are necessary,” says Hawkins.
Stellar’s expertise in building design and construction provides an opportunity for the firm to target PdM with other critical assets, including industrial refrigeration and mechanical utilities space, compressors, condensers, piping, valves, evaporators, pumps, fans and vessels. The firm offers a range of PdM and PM programs including predictive solutions, condition-based monitoring, inspection, mechanical integrity, rotating assets and vibration or oil analysis.
Lubricants should also be included in a preventive program, because their proper use can reduce downtime and increase profits, says Toby Porter, food market manager at Klüber Lubrication.
For instance, a Klüber customer used its PdM program to look at conveyor chain life trends. The beverage plant routinely shut its machinery down to replace chains experiencing excessive elongation. PdM revealed improper lubrication caused the chains of a conveying unit to stretch one to two inches, which led to their out-of-spec performance and eventual failure sooner than anticipated.
Having a predictive maintenance program geared around lubrication management not only resulted in the use of a more appropriate lubricant, it showed the benefits higher-performance oil provides. “Lubricants typically make up only 1 percent of a company’s total operating costs, but the oil chosen can have a significant impact on the cost of equipment, labor and energy,” says Porter.
Here’s an example of what Porter means: A food plant that amassed 8,000 hours of operating time per year was able to reallocate almost 1,500 man-hours from lubrication maintenance by switching to a specialty synthetic lubricant that extended the interval between oil changes. The switch also trimmed the frequency and amount of used oil the plant generated and disposed of, and lowered the processor’s energy costs due to the oil’s lower coefficient of friction.
For more information:
Terrance Fisher, SKF USA, 585-314-2020, email@example.com
Jim Freaner, ATS, 309-693-5926, firstname.lastname@example.org
Chris Hanssen, JBT FoodTech, 507-664-7319, email@example.com
Steve Hawkins, Stellar, 904-260-2900, firstname.lastname@example.org
John Kravontka, Fuss & O’Neill Manufacturing Solutions, email@example.com
Katie Moore, GE Intelligent Platforms, 206-550-0549, firstname.lastname@example.org
Toby Porter, Klüber Lubrication, 800-447-2238, email@example.com
Matt Ruth, Avanceon, 610-458-8700 ext. 274, firstname.lastname@example.org