Transcription

Predictive maintenance andthe smart factoryPredictive maintenance connects machines to reliabilityprofessionals through the power of the smart factory

Predictive maintenance and the smart factoryTraditionally, maintenance professionals have combined many techniques,both quantitative and qualitative, in an effort to predict impendingfailures and mitigate downtime in their manufacturing facilities. Predictivemaintenance offers them the potential to optimize maintenance tasks inreal time, maximizing the useful life of their equipment while still avoidingdisruption to operations.IntroductionMaintenance and reliability professionals in themanufacturing industry face a number of challenges, butthe goal of any maintenance organization is always thesame: to maximize asset availability. In this article, we’llfocus on fixed assets, the machines on the shop floorthat turn raw material into finished goods.Today, poor maintenance strategies can reduce a plant’soverall productive capacity by 5 to 20 percent.1 Recentstudies also show that unplanned downtime is costingindustrial manufacturers an estimated 50 billion eachyear.2 This begs the question, “How often should amachine be taken offline to be serviced?” Traditionally,this dilemma forced most maintenance organizationsinto a trade-off situation where they had to choosebetween maximizing the useful life of a part at the riskof machine downtime (run-to-failure) or attemptingto maximize uptime through early replacementof potentially good parts (time-based preventivemaintenance), which has been demonstrated to beineffective for most equipment components.Oftentimes, maximum utilization of tooling or machinecomponents can be achieved by running them until they02fail. But this can lead to catastrophic machine damage asparts begin to vibrate, overheat, and break, reinforcingthe old adage “pay me now or pay me later.” And, whilerun-to-failure may be an acceptable approach for someassets, it still needs to be understood that unplanneddowntime is almost always more expensive and timeconsuming to correct. Conversely, you might considermore frequent replacement of parts and servicing ofequipment. But this can not only increase replacementcosts over time, it can also increase planned downtimeand disruption to operations.Spare-parts management presents a similar challengethat can feel like a constant balancing act. With limitedbudgets, maintenance professionals must evaluatewhich parts they’ll need and when to procure them. Ifthe spare is not on hand or on order when it’s needed,the downtime of an asset can be anywhere from daysto weeks—or even months—while waiting for thereplacement part. This typically leads to the buildup ofspares inventory, which not only ties up working capital,but also increases the risk of excess and obsolescencethat erodes the bottom line.

Predictive maintenance and the smart factoryBreaking the tradeoff: Leveraging the powerof the smart factoryPredictive maintenance (PdM) aims to break thesetradeoffs by empowering companies to maximizethe useful life of their parts while avoiding unplanneddowntime and minimizing planned downtime. With theadvent of Industry 4.0 for manufacturing, companies areable to leverage new technologies in order to monitorand gain deeper insight into their operations in realtime, turning a typical manufacturing facility into a smartfactory. Simply put, a smart factory is one equipped withtechnology that enables machine-to-machine (M2M) andmachine-to-human (M2H) communication in tandemwith analytical and cognitive technologies so thatdecisions are made correctly and on time.PdM (which has been around for many years now)utilizes data from various sources, such as criticalequipment sensors, enterprise resource planning (ERP)systems, computerized maintenance managementsystems (CMMS), and production data. Smart factorymanagement systems couple this data with advancedprediction models and analytical tools to predict failuresand address them proactively. Additionally, over time,new machine-learning technology can increase theaccuracy of the predictive algorithms, leading to evenbetter performance.In contrast, traditional preventative maintenance (PM)programs often require very time-consuming, manualdata crunching and analysis to gain any real insightsfrom the data being collected. While many have hadsome success with these strategies, they typically relyheavily on “tribal knowledge” estimates or requirein-depth knowledge and analysis of each individual pieceof equipment on an ongoing basis to stay accurate.In line with the maintenance goal of maximizing machineavailability, Deloitte identified two main businessobjectives of all manufacturing companies operating inthe Industry 4.0 era:31. Operating the business2. Growing the businessyou consider the number of person-hours spentperforming routine machine inspections that don’t leadto any key findings or trigger a work order, in addition tothe effort spent firefighting unplanned downtime, thecase for PdM starts to become clearer. PdM technologiescan pull data from multiple sources and legacy systemsin order to provide real-time advanced insights, allowingcomputer systems to do the legwork so maintenancemanagers can deploy their resources more efficientlyand effectively.Getting to predictive maintenance: An explorationof the technologiesMaintenance organizations across industries are, bydesign or default, at different stages of maturity. Somemay be running scheduled maintenance checks basedon estimates or OEM recommendations, while othersmay utilize statistics-based programs individuallytailored to each fixed asset. Still, others—particularlyin the aerospace and energy sectors—are alreadyemploying continuous monitoring technologies of theirassets, but may only be monitoring the outputs of thedata, rather than leveraging advanced predictive models.As with anything, there are steps to take on the journeytoward reliability optimization, beginning with some ofthe basics of preventive maintenance and reliabilitycentered maintenance, while simultaneously pilotingPdM with one or two well-suited assets. Prime assetsfor one of these pilots should be highly integral tooperations and must fail with some regularity in order tocreate baseline predictive algorithms.Now, the idea of PdM sounds enticing. But how does itwork? Many of the technologies that make up a smartfactory are not necessarily new, but have become muchmore affordable, more robust, more advanced, andintegrated for business use. Computing, storage, andnetwork bandwidth all are now available at fractions ofthe cost compared to just 20 years ago,4 making pilotingand scaling financially feasible.Let’s explore some of the technologies that make up asmart factory and make PdM possible.While growing the business focuses on top-line growth,operating the business often seeks to cut costs. When03

Predictive maintenance and the smart factoryInternet of ThingsThe Internet of Things (IoT) is perhaps the biggestpiece of the PdM puzzle. The internet as we know ithas connected your laptop and mobile device to largeserver farms full of website data coded in HTML. IoT issimilar, but the data is created in a continuous streamfrom your assets to private enterprise servers. IoTtranslates physical actions from machines into digitalsignals using sensors such as temperature, vibration,or conductivity. Data can also be streamed from othersources, such as a machine’s programmable logiccontroller (PLC), manufacturing execution system (MES)terminals, CMMS, or even an ERP system. IoT completesthe first half of the physical-to-digital-to-physical (P-D-P)loop (Figure 1 below). This smart factory concept wasintroduced in Deloitte’s discussion on The Rise of theDigital Supply Network. Once the physical actions havebeen translated into digital signals via sensors, theyare processed, aggregated, and analyzed. With theaffordability of bandwidth and storage, massive amountsof data can be transmitted to give not only a full pictureof assets in a single plant, but of an entireproduction network.Figure 1: Physical-to-digital-to-physical loop and related technologies2. Analyze and visualizeMachines talk to each otherto share information, allowingfor advanced analytics andvisualizations of real-timedata from multiple sources1PHYSICAL1. Establish a digital recordCapture information fromthe physical world to create adigital record of the physicaloperation and supply network2DIGITAL33. Generate movementApply algorithms and automationto translate decisions andactions from the digital world intomovements in the physical worldSource: Center for Integrated Research. Deloitte University Press dupress.deloitte.com04

Predictive maintenance and the smart factoryAnalytics and visualizationThe second step in the P-D-P loop is to analyze andvisualize the digital signals using advanced analytics andpredictive algorithms. Advanced business intelligence(BI) tools are no longer only for data scientists. Manyanalytics platforms are beginning to incorporatehigh-level solutions for unstructured data, cognitivetechnologies, machine learning, and visualization.Operations analysts, who are more in-touch with themanufacturing processes, can easily create dashboardsusing modern APIs (application program interfaces)created specifically for the everyday user.Another trend is data moving back to the edge. Similar tothe lean technique of storing tooling at the point-of-use,data computation will be done at the “edge,” meaningit’s processed at the machine where it’s generated.Insights can be pushed directly to machine operatorsand maintenance technicians. As data is beginningto approach the zettabytes (that’s 1021 bytes), edgecomputing reduces the overall burden on a computernetwork by distributing some of the processing work toa network’s outer nodes to alleviate core network trafficand improve application performance.5system, check the ERP system for spares on hand,and automatically create a purchase request for anyadditional parts required. Then, the maintenancemanager only has to approve the items in the workflowand dispatch the appropriate technician, all automatedand prior to unplanned downtime.Potential benefits:The challenges at first glance might seem steepto overcome. However, the benefits of digitaltransformation far outweigh the risks. Thesebenefits include:6 Material cost savings (5-10 percent in operationsand MRO material spend) Reduced inventory carrying costs Increased equipment uptime and availability(10-20 percent) Reduced maintenance planning time(20-50 percent) Reduced overall maintenance costs(5-10 percent) Improved HS&E complianceClosing the physical-to-digital-to-physical loopFinally, after the signals have been processed, analyzed,and visualized, it’s time to turn those insights back intophysical action. In some cases, the digital conclusionsdrawn may instruct robots or machines to alter theirfunctions. In other cases, maintenance alerts will spura technician into action. Consider a situation wherethe predictive algorithms would trigger the creationof a maintenance work-order in the company’s CMMS Less time spent on brute-force information extractionand validation More time spent on data-driven problem solving Clear linkages to initiatives, performance,and accountability More confidence in data and information leading toownership of decisions05

Predictive maintenance and the smart factoryBuilding the foundation: Seven maintenancepillars for successTechnology alone cannot get you there—you alsoneed to focus on process and organizational changes.Successful maintenance organizations should be ableto deploy all appropriate resources (human resources,technical instructions/reference materials, spareparts, and tooling) when and where they are neededin support of operations. To accomplish this, there area number of key areas that need to be addressed. Alltoo frequently, companies tend to spend their time andmoney on major technology enhancements, such asmore robust computer-managed maintenance systems(CMMS) or reliability software, without first applyingenough focus on some of the more fundamentalelements of the organization. It should be understoodthat there is no “silver bullet” computerized solutionwhich will alleviate the need for some initial ground work,including determining exactly what type of maintenanceorganization and methodology makes sense for aparticular operation. In order to optimize a maintenanceorganization, there are seven primary focus areas thatshould be considered (see Figure 2).Maintenance strategy and processes are the core elements for anysuccessful maintenance organization. And it’s important to notehere that while technology is a key enabler (and the primary focus ofthis paper), it’s only one of the seven pillars for success. Without thefundamental building blocks in place, investment in technology willlikely never yield the desired results. It should also be understoodthat not all companies require the same level of reliability from theirassets. A good place to start is to assess your organization’s missionrequirements and maintenance program maturity. Ask yourselfsome of the following questions:Figure 2: Maintenance pillars for success What data do we already have that isn’t beingused effectively?Technology How reliable do our assets need to be? What are ouravailability targets? Do our technicians have the right skills to perform thework required? Do we have the right spare parts in the right place at theright time? Are our processes well-documented, accessible,and useful? Do we have the right tools for the job? How do we determine when it’s time to replace an asset ratherthan maintain it? Have we identified the critical assets in ourproduction system?People Are there some critical assets that would benefitfrom a PdM pilot? What is the value of PdM across our entire enterprise?EquipmentMaintenancestrategy andprocessesTools06Doc