ANALYST REPORT

INDUSTRIAL IoTHow manufacturers will optimise performance

and maximise opportunities

IN ASSOCIATION WITH:

CONTENTS

ANALYST REPORT30 Optimise performance in manufacturing with Industrial IoT

31 Monitoring, remote control and management

32 Efficiency, productivity and automation

33 Operational data spaces and closed loop efficiency models

34 Predictive maintenance, delivering the service and digital twins

35 The future of servitization

36 External data, data monetisation and operational data spaces

37 Conclusion – The future of smart manufacturing

GROWTH IN GLOBAL IoTCONNECTIONS IN CONNECTEDINDUSTRIES 2015-2025

30FIVE LEVELS OF INDUSTRIALIoT BENEFITS ANDOPPORTUNITIES

31

29IoT Now - February / March 2017

Global manufacturing faces an increasing set of challengesin growth, innovation and competition. Globally, grossdomestic product (GDP) was at US$30.69 trillion in 1995.By 2014, global GDP reached a new peak at US$78 trillion.As a percentage of global GDP, manufacturing dropped itsshare from 21.4% in 1995 to 14.7% in 2014. In real terms,manufacturing grew, and yet the rate of growth wassubstantially less than that of the global economy.

Industrial IoT (IIoT) has accelerated the pace of innovationand manufacturers are pushed to explore and implementnew technologies in IoT while maintaining existing systemsand architectures, creating tensions between operationaltechnologies, information technology and IoTimplementations. Differences between regions andmarkets do exist, and globalisation has presented greatercompetition between markets than ever before. Commonto all manufacturers are the continuous needs to reducecosts, improve performance and deliver quality tocustomers. Combined growth, innovation and competitionhave the manufacturing industry looking to industrial IoTas a way forward, and as shared in this report, exciting andinnovative IoT solutions are reinventing manufacturing.

Growth of connections inconnected industryConnected industry covers a range of industries andapplications, and includes several application groupsdefined by Machina Research. Manufacturing andprocessing, supply chain, warehousing and storage, andextractive industries are some of the application groups.

From Machina Research’s global forecasts, it is noticeablethat established manufacturing regions such as Europeand the US continue their innovation paths with IoT at asteady pace of 12% per year in IoT connections. Incontrast, emerging Asia-Pacific with the lion’s share in

China, grows at 15% year-on-year, not yet eclipsing thecombined number of connections for Europe and the USbut reducing the gap from 43% to 29% in the ten-yearperiod of 2015-2025, showing a real investment byemerging industries in the new technologies.

Figure 1: Growth in global IoT connections in connectedindustry 2015-2025 [Source: Machina Research, 2016]

Benefits and opportunities fromindustrial IoTIndustrial IoT offers manufacturing an emerging anddeveloping set of benefits and opportunities. Five levelsare identified: monitoring, remote control andmanagement, efficiency and productivity, automation andoperational data spaces. Figure 2 illustrates therelationship and progression between the five levels. Eachlevel involves a complete IoT architecture andinfrastructure of connected devices, networks, platformsand applications, generating real-time data from theconnected devices, and enabling the processing andanalysis of the data for different outcomes.

30 IoT Now - February / March 2017

ANALYST REPORT

Optimise performance inmanufacturing with industrial IoT

The author is Emil Berthelsen, principalanalyst at Machina Research

▼

31IoT Now - February / March 2017

Figure 2: Five levels of Industrial IoT benefits andopportunities [Source: Machina Research. 2016]

MonitoringIndustrial IoT has a long history with telemetry andsupervisory control and data acquisition (SCADA)solutions, delivering the earliest examples of connecteddevices to monitor usage, condition, and performance,and through alerts and messages, improve the overallmanagement of different operational processes. This couldrange from monitoring production lines, to drillingequipment at oil and gas sites, to being embedded indefence and space vehicles, and in monitoring longstretches of pipelines, networks and remote weatherstations.

Basically, telemetry and SCADA solutions have enhancedoperational processes by enabling enterprises to monitorremote locations and equipment, reducing costs on 24x7manning and/or receiving more timely information onprocess failures or poor performance. Being able tomonitor processes in near real-time has also added to theoverall operational performance of such machines asconstruction equipment, transportation and ships.

One fundamental development that did need to take placeto enable remote monitoring to work to its full extent waswireless wide area network (WAN) communications.Without continued developments in quality and coverage

of radio frequency including satellite and cellulartechnologies, the value of remote monitoring would nothave been as great as that now enjoyed.

Remote control and managementA next development step from being able to monitorconnected devices is to enable actuation, which is remotecontrol and management. Achieving this sounds like astrange task until you begin to explore the requirements interms of the connected device itself – and itsconfiguration, the type of connectivity used, andultimately, how to manage thousands and thousands ofcommands to different devices at the same time.

Within the configuration of remote control andmanagement are the steps of receiving and analysing thedata off the connected devices, defining immediateactions or commands, and looping these back to thedevices, or other devices in the wider solution. This could,for example, be information from a production facilitymeasuring the level of humidity in the manufacturingprocess and potentially issuing a command to turn on theventilation systems for a given amount of time to reach adefined setting level. Another example would be an officebuilding automation system which provides temperaturereadings from the office building and automaticallyinitiates the air conditioning systems and window blinds toreach a cooler temperature with less sun impacting thetemperature levels.

Remote control and management has meant furtherdevelopments in associated technologies. At the devicelevel, actuation capabilities have become an importantfeature of M2M and IoT solutions. Previously, moduleswere mainly configured with a sensor, very limitedprocessing capabilities and a wireless connectivity solutionand antenna, sending alerts and messages. With therequirements of remote control and management,particularly including actuation, connected devices nowrequire additional processing and actuation features aswell as two-way communications in the connectivitytechnology.

Another feature of remote management that has had ▼

significant implications for industrial IoT, connecteddevices and their associated lifecycles are over-the-airupdates. For many connected devices with lifecycles inthe five to 12 year range, and more importantly, installed inremote and potentially hard to access locations, remotemanagement including feature updates has become a veryimportant attribute in IoT. With two-way communications,connected devices have had to become significantly moreflexible and open in their structures, enabling futureconfiguration development and as will be seen later,opening the path towards fog computing and edgeanalytics for automation.

Industrial IoT has moved the goalposts from passivemonitoring to more of a direct interaction between theconnected devices and the enabling IoT platforms.

Efficiency and productivityFrom monitoring and remote control and management,the next set of opportunities from connected devices inreal-time was managed through the data. M2Mapplications were designed with narrow functionalityobjectives, and connected devices were part of the designarchitecture to capture and generate the data from theenvironment for the specific applications. In IoT, followinga growing amount of data generated and transmitted byconnected devices in real-time, the opportunities toanalyse this data further and potentially for purposesother than the M2M or IoT applications became arecognised opportunity. Data from connected devices wasused for the designed M2M/IoT application but soonbecame an asset which could be used for otherapplications, combined with other data sets, andultimately, a monetisable asset by itself.

With multiple data sets becoming available from variousconnected machines, enterprises and their data scientistshave been able to develop substantially more accuratepictures of the equipment including wider rangingprocessing and analysis of the data. As this data hasincreased in scale, speed and structure, and as enterpriseshave developed and applied advanced analytical tools andmachine learning, the outcomes of data processing havebecome even more promising.

New insights based on for example predictivemaintenance, fraud detection and condition monitoringhave moved the goalposts from describing how things areto what could happen, and allow for enterprises to takepreventive action. This preventive action is turn can saveindustries millions of dollars in terms of minimised

operational disruptions, improved machine efficiency andproductivity and early detection of faults and fraud.Consider the following examples which have becomecornerstone applications in industrial IoT. Within mostindustries, maintenance routines are either based ondefined schedules or some method of condition-basedmonitoring. Operators of, for example, wind turbines orrolling stock would carry out maintenance routinesaccording to set schedules which in many cases wereirrelevant to whether the actual equipment required theservice routine or not.

While routines have helped structure unplanneddisruptions into predicted activities, with connectedmachinery, enterprises monitor in real-time the operationalperformance of machines, and with the assistance ofadvanced analytics, predict with greater degrees ofaccuracy when maintenance is required. In particular, theability to aggregate and analyse data from multiplesensors such as temperature, vibration, noise and imageshave enhanced the ability to detect future faults withgreater accuracy. Predictive maintenance has become asignificant and substantial new service available toenterprises from their connected solutions.

As an important side note, predictive maintenanceservices have become a crucial example of howequipment manufacturers have been able to extend theirportfolio from equipment sales, to equipment plus serviceas devices are now connected, providing real-time data.This has unlocked substantial opportunities for equipmentmanufacturers to enhance their product and serviceportfolios, become service providers, and establish newand innovative relationships with their customers. Thisconcept of servitisation, augmenting product sales with newservices based on the data from connected equipment, isone that will be explored further in this report.

What applications such as predictive maintenance, frauddetection and condition monitoring bring to IoT andenterprises is a greater efficiency and productivity tooperational performance. This development is a stepfurther on from monitoring and remote control andmanagement, and involves additional tools such asadvanced analytics and machine learning, each in turn alsoimproving with more and more domain expertisebecoming involved in the development of these tools.

AutomationManufacturing industries have followed developmentsfrom early M2M to the current stages of industrial IoT

32 IoT Now - February / March 2017

ANALYST REPORT

▼

33IoT Now - February / March 2017

including monitoring, remote control and managementand efficiencies and productivity, and, as one executiveexpressed, the next goal is to remove “the security risks inoperations by operators.” In high volume, high speedproduction environments, real-time and nearinstantaneous decision-making has become a criticalcomponent in manufacturing environments. Withanalytical tools able to ingest, process and analysesubstantial amounts of data in real-time, the goal of fullautomation within operations is coming closer and closerwithin reach. Based on operational performance data andprescriptive analytics, and driven by such significantinitiatives as Industrie 4.0 in Germany, Smart Manufacturingin the US, and many innovation centres in Japan, Europeand China, the goal is to achieve operational efficiency andproductivity through automation.

Operational data spaces and closedloop efficiency models All four levels of monitoring, remote control andmanagement, efficiency and productivity, and automationshare a common set of attributes. All these operationalimprovements and enhancements are based on enterprisedata, generated by connected devices, and fed directlyand immediately back into the operational loops in aclosed loop efficiency model. It’s data from enterpriseequipment, used for enterprise operational improvements,and delivering operational results. Yes, in monitoring andremote control and management, the solution has movedon from what are termed steady state models such asthose designed to monitor and achieve an operationalsteady state although time lags in decision-making maytake place. In efficiency, productivity and automation thesolutions have shifted towards a what-if/when anticipatorystate, analysing the data, predicting future operationalscenarios and either augmenting existing decision-makingprocesses with real-time data or looping this informationdirectly back to the operational processes, and adjustingthe process. This becomes the move from descriptive topredictive and prescriptive action for manufacturers basedon the data from the connected devices, and begins toshow the new benefits and opportunities from internal,enterprise data.

At the next level, operational data spaces, manufacturersshould begin to explore the boundaries beyond enterprisedata, and understand how aggregated and augmenteddata, combining internal with external data delivers evenfurther and additional value to the enterprises. Beforethen, it is worth identifying and sharing many of the other

ways enterprise data has been exploited, particularly byother supporting processes within the business, and in themore general fields of digital twins and servitisation.

Additional opportunities enabled by enterprise dataDevices producing real-time data about the status,condition and performance of connected equipment hasunlocked new, innovative and additional opportunities forenterprises in other supporting manufacturing processes.Three supporting processes, core in their own rights, aresupply chain, product development and management,and maintenance.

Supply chain improvementsOne of the critical and core processes to develop moreefficient manufacturing processes is that of the supplychain. The close alignment and management of supplychains and the manufacturing processes enable enterprisesto optimise the utilisation of machinery and plan resourcesin excruciating detail, including production schedules andruns. Managing supply chains has been at the centre ofimprovement initiatives in manufacturing for severaldecades – see Kanban or Just-in-Time manufacturing, andwith emerging IoT technologies, the level and quality ofinformation from the origin of raw materials to theproduction facilities has improved significantly.

To minimise costs, for example, raw materials held in theprocess, manufacturers have looked to balance the timelywarehousing and stock control of raw materials with theproduction requirements on the manufacturing floor.Previously planned around set delivery schedules withbuffer stocks, closely monitored and managed supplychains with IoT enable enterprises to improve theseprocesses substantially, following every step of the rawmaterials from origin to production. Industrial IoT hasenabled enterprises to understand both the location ofraw materials at any given point in time but equallyimportant for many industries such as food processingand pharmaceutical industries, the conditions under whichthe raw materials have been transported.

As an example, manufacturers able to monitortransportation conditions of raw materials requiringminimum and maximum temperature or humidityconditions, can detect well in advance of the productionprocesses if anything may have gone wrong duringtransportation – increased humidity for example, andavoid further unnecessary production costs with qualityfailures later in the processes. ▼

34 IoT Now - February / March 2017

ANALYST REPORT

Equally, given the accuracy of location data andcalculations of arrival times, manufacturers have been ableto work against more accurate production schedules andquickly align production rates where required, to keep linebalancing at an optimum. For substantially largerenterprises, improved supply chain management has asignificant impact on efficiencies and cost control ofmanufacturing processes. It delivers smoother and moreefficient production runs, fewer unnecessary productiondisruptions due to material shortage, and avoids wastedproduction runs with inferior materials impacted duringtransportation.

Transformation of product development and managementAnother significant development in industrial IoT is theimpact on product development and management.Previous product development processes includecustomer feedback about the use and performance ofproducts once launched in the field. These feedbackprocesses were either at set service intervals or whenissues arose with the product or sometimes never.

As products have become connected, transmitting real-time data about their status, condition and performance,product design and engineers can monitor productsthroughout their entire lifecycle. Through design stages toactivation stages and to actual implementations orcustomer ownership, products are now monitored, andinstantaneous feedback about product condition andperformance is being captured, processed and analysed.

For example, manufacturers of tires such as Continentaland Pirelli have implemented for several high-end tireranges, the connected tire, allowing manufacturers tomonitor tire condition and performance, and include anyimmediate findings into their production processes.Similarly, automotive manufacturers have followed suit,constantly evaluating, assessing and controlling forexample engine performance, fuel mix, brakes, power,drivetrain, and a host of other activities.

Vendors have recognised this development quite early onin IoT marketplaces and offer, in addition to theirtraditional software based design tools, the ability forenterprises in general but more specifically for themanufacturing industry, to launch connected products andbenefit immediately from the real-time data off thedevices. These benefits have also been shared withcustomers, as this report will explore in the section onservitisation.

Predictive maintenance, delivering theservice and digital twinsPredictive maintenance has been heralded as one of thesignificant opportunities emerging from industrial IoT forthe manufacturing industry, and it’s easy to understandwhy given the cost of machinery and efforts to maximiseutilisation without damaging the equipment.

Efficient maintenance routines have always been key tothe successful management of equipment whether youare in any production industry, transport industry, or anyindustry where critical assets enable service delivery. Fordecades, maintenance solutions have taken variousapproaches including planned maintenance schedules,visual inspections, condition-based monitoring systemsand reliability solutions. The emergence of industrial IoThas added a substantial factor in all the above approachesin terms of real-time data, and has even gone several stepsfurther by including new analytical processes for potentialfault and failure detection, or in other words, predictivemaintenance.

In predictive maintenance, enterprises are accurately ableto analyse the status, condition and performance ofequipment through multiple sensor data sets including thetraditional performance data of production speeds andoutput and more recently, aggregating these data setswith noise, vibration, humidity, temperature and evenvisual data inputs. Combined, multiple data sets areanalysed and provide substantially improved insights forpredictive maintenance purposes.

Predictive maintenance becomes a substantial benefit andopportunity for the running and management ofmachinery in the manufacturing industry. What industrialIoT in addition delivers to maintenance is the actual way ofservicing machines. Previously, service engineers will haveapproached maintenance and repair tasks with diagnosticresults and potential recommendations for problemresolution. Their expertise helped them through the task.With industrial IoT, service engineers are geared with moreadvanced information available, and two new gadgets,tools and applications for service engineers come into play– augmented reality and the digital twin.

How augmented reality adds significantvalue to the work of service engineersService engineers equipped with augmented reality (AR)glasses and tablets and associated engineering ▼

35IoT Now - February / March 2017

applications have visual diagrams of machine equipmentsuperimposed on actual equipment, and provide real-timeguides and recommendations as to how to completerepairs and perform maintenance tasks. For serviceengineers, AR glasses provide additional data for greaterengineering efficiency, and particularly for lessexperienced engineers, AR glasses and the associatedapplication may provide invaluable service task assets. Inan example shared by PTC and one of its customers,Getinge, the company showed how medical equipmentcould be easily serviced by on-site engineers with the newtechnology, and how more complex maintenance taskscan be improved with the technology.

How digital twins will become acommon design feature of the futureA digital twin is the computerised version of a physicalasset, providing a detailed software-based visualisationand composition of the actual piece of equipment orproduct. Product designers have worked with earlierconcepts of digital twins in the form of computer-aideddesign (CAD) and software simulations. With theemergence of industrial IoT and connected products,these digital twins have been brought to life from staticcomputer-aided designs to visual displays sharing real-time data from the products as well as simulated effectswhich the data may indicate. Looped back, the digital twinbecomes both the source for invaluable information andinsights for product engineers as well as the tool forservice engineers to work against in augmented reality.

For manufacturers, the record of product design andservice management routines are becoming based moreand more around digital twins, and with the conceptdeveloping together with maintenance, productengineering and customer experiences, digital twins are anunquestionable asset of the future.

The future of servitisationIndustrial IoT has been described as a disruptive marketforce, and servitisation is at the roots of this thinking,making this a critical opportunity for original equipmentmanufacturers (OEMs), service and solution providers, andapplication developers. In fact, servitisation unlockssignificant opportunities for the entire IoT ecosystemincluding the manufacturing industry. With connectedproducts and services, manufacturers can radicallytransform their business models from product-oriented

business models to service-oriented models withsubstantial impacts on the business such as financialmodels, customer lifecycles and skills and resources. Allthese changes form part of the digital transformation ofenterprises, and form an integral part of the growth,innovation and competition challenges faced by themanufacturing industry. What are the changes in financialmodels, customer lifecycles and skills and resources forenterprises in Industrial IoT?

Moving from capital expenditure and operatingexpenditure financial modelsManufacturers have traditionally focused on product-oriented business models. Here, transactions around thesale of products have been the dominant business model,and in heavy manufacturing, for example, the sale of heavyequipment has been based on capital expenditure modelsfor the equipment and the purchasing enterprise. Withindustrial IoT and connected equipment, more and moremanufacturers are transforming their business models tomore a service-oriented model based on a service chargeon equipment usage rather than asset sale. What thismeans for manufacturers is a realignment of financialmodels from capex-driven to opex-driven models, and forenterprises procuring services, the opportunity to avoidsteep upfront capital expenses, and pay for the equipmentas it is used. For those manufacturers who have yet toadopt the new business models, competition has begun toappear quite intense, and it is worth remembering, that allthis is only enabled through industrial IoT.

New customer lifecyclesMoving from a product-oriented business model to aservice-oriented business model also means a significantchange in the relations between the manufacturer and thecustomer. From a single, across the counter transaction,the service-based model opens an extended andimportant customer relationship which manymanufacturers will not have dealt with before. While thisunlocks opportunities for additional sales of services, italso places new responsibilities on the manufacturer interms of customer service support and maintainingongoing customer relations. Organisational features whichmany manufacturing businesses will need to develop andimplement.

The new customer lifecycle challenges for manufacturerswill evolve in parallel as part of the digital twin and ▼

36 IoT Now - February / March 2017

ANALYST REPORT

predictive maintenance services which manufacturers willbegin to offer their customers. In this emerging model ofinnovation, competition becomes increasingly moredifficult for latecomers to the market as servitisationbuilds closer relations between manufacturers and theirend customers.

Skills and resources in the industryServitisation and industrial IoT are two disruptive forceswhich executives in the manufacturing industry need toprepare for and engage in. The do nothing approach is notan option. And in transforming the business, executiveswill need to identify and develop their skills and resourcesas part of the digital transformation. This includesresources in product engineering, sales and customerservices as well as newer skills and resources in businessand data analysis and business development.

The benefits and opportunities for manufacturers forgrowth and innovation are significant with servitisationand industrial IoT but as with other disruptive forces, themarket does not stand still, and new competitive playersand markets will emerge with the changes. Thiscompetition brings us to the next and future developmentof IoT data, moving from internal monetisation to externaldata monetisation based on operational data spaces.

External data, data monetisation andoperational data spacesMonetisation of data for enterprises has followed a path ofinternal data monetisation for internal operationalobjectives, and with experience and the right datamanagement tools in place, enterprises have cautiouslystarted to explore external data monetisation, ultimatelyaimed at producing new revenues streams for the

business. These remain early days for monetisation goalswith internal and external data, and for most enterprises,especially those within the manufacturing industries,internal data monetisation remains the priority. Figure 3illustrates the four quadrants in which internal and externaldata can be monetised. The following provides a quickdescription of each quadrant:

Quadrant one: external data for internal monetisationobjectives. Enterprises have a history of merging externalmarketing data sources with minimal customer data,building marketing strategies, defining new customersegments and create new product and servicepropositions.

Quadrant two: internal data for internal monetisationobjectives. Enterprises have long tapped into historicalenterprise data sources such as key performanceindicators (KPIs) for operational improvement reasons,and with increasing amounts of data emerging fromindustrial IoT, enterprises are quick to identify operationalimprovement opportunities such as predictivemaintenance.

Quadrant three: internal data for external monetisationobjectives. The focus on internal monetisation remains thepriority for most enterprises. The opportunities fromselling internal data, anonymised and pseudonymised aretightly weighed against the potential and perceived costsof data privacy, governance and security. As technologiesimprove, and as more defined data communities orSubnets of Things are established with the appropriateand suitable levels of authority and security in terms ofdata sharing, enterprises will start to engage. Similar datasharing models have been seen in other industries such asthe airline and shipping industries, the hotel industry andmore openly, in many scientific fields. ▼

ANALYST REPORT

37IoT Now - February / March 2017

Quadrant four: external data for external monetisationobjectives. The open data market remains a distantprospect for the time being. Enterprises remain focusedon internal monetisation paths, and will continue toexplore and follow developments in data privacy andsecurity. Once markets fully accept data-sharing and theappropriate regulatory frameworks for data exchange andoperational data spaces are established, this opportunityarea may be explored but that remains some time away.

Figure 3: Internal and external data monetisation paths forenterprises [Source: Machina Research, 2016]

The future of smart manufacturingGrowth, innovation and competition are three of the mainchallenges faced by industry. These have been continuouschallenges for any manufacturing executive, and yet, thepace of change and scale of disruption with the Internet ofThings and servitisation has led to numerous articles andpresentations mentioning the impacts of the nextIndustrial Revolution.

Growth of connected devices and data is no longersomething for the future. It is here. The behemoths of IoTsuch as IBM, GE, SAP, Microsoft and AWS are quotingmillions of connections being enabled each year, and thesame players are at the forefront of the innovation in dataand analytics. Manufacturers are not immune to thesedevelopments, and will be adding to these connectionsand developing the tools for data and analytics.

In innovation, manufacturers are reinventing andredeveloping their businesses, and improving operationalperformance, customer experiences and new revenuestreams. Transforming the business from a product-oriented industry to a service-oriented industry isunderway.

Finally, competition. As manufacturers are reinventingthemselves, they are not only unlocking new opportunitiesbut establishing and entering new competitive marketswith players extending their capabilities. This has led tothe development of new ecosystems and approaches inpartnerships and collaborations, and manufacturers willquickly recognise that yesterday’s single enterpriseempires will be made up of conglomerates andpartnerships of tomorrow.

Machina Research provides market intelligence andstrategic insight on the newly emerging Internet ofThings (IoT), Machine-to-Machine (M2M) and big dataopportunities.

The Internet of Things is the single most importanttechnology trend today. IoT technologies are alreadyenabling new and innovative business opportunities,proving a significant disruptor of traditional businessmodels and processes. It is front-of-mind for manycorporate management teams as well as the myriad oftechnology vendors that support and supply them.

Staffed by industry veterans, we provide marketintelligence and strategic insight to help our clientsmaximise opportunities from these rapidly emergingmarkets. If your company is a mobile networkoperator, device vendor, infrastructure vendor, serviceprovider or potential end user in the IoT, M2M, or bigdata space, we can help.www.machinaresearch.com