Simplifying smart manufacturing concepts

(1)

IN

DEGREE PROJECT MECHANICAL ENGINEERING, SECOND CYCLE, 30 CREDITS

STOCKHOLM SWEDEN 2018,

Simplifying smart

manufacturing concepts

Commercialization of smart manufacturing concepts to help small and medium-sized enterprises

BHARAT SHARMA

PRAVEEN NATARAJAN

KTH ROYAL INSTITUTE OF TECHNOLOGY

SCHOOL OF INDUSTRIAL ENGINEERING AND MANAGEMENT

(2)

Simplifying smart manufacturing concepts

Commercialization of smart manufacturing concepts to help small and medium-sized enterprises

Bharat Sharma Praveen Natarajan

MG203X degree project in Production Engineering and Management KTH Royal Institute of Technology

Stockholm 2018

(3)

Abstract

Smart manufacturing is a concept that has taken over the manufacturing sector. The definition of smart manufacturing or the German concept Industry 4.0 has not been clearly stated and the transition from Industry 3.0 to 4.0 is still debated. Large companies have started investing a lot of money to research on concepts like Industrial Internet of Things (IIoT), advanced data analytics and cloud computing, that compliment smart manufacturing to fit their manufacturing setup and hence the solutions are not standardized. They have still managed to improve their production systems to make it more flexible, productive, efficient and customer-centric by using the concepts of smart manufacturing. Small and medium-sized enterprises are challenged by problems faced such as low investments, insufficient infrastructure, inefficient production lines and limited time to research on new technology. This thesis studies the concept of Industry 4.0 or smart manufacturing and tries to create a transition from Industry 3.0 concepts to 4.0 concepts by using simple and easy solutions. The main aim of the thesis is to create a business plan to commercialize the simple solutions so that small and medium-sized enterprises could start implementing the smart manufacturing concepts in a simple way. Two concepts namely machine vision system and Bluetooth Low Energy technology were studied, and a product structure was developed. Real- world cases were developed and tested out with the product. Conclusions were made as to how the products and services would benefit the customer and what problems it addresses on a shop- floor setting. A structured business plan to commercialize the product was developed with a customer-centric approach, that helps create a transition from Industry 3.0 to 4.0 concepts as well as provide cost-effective and easy solutions to small and medium-sized enterprises.

Validation of the concepts were done using concepts like Bluetooth Low Energy (BLE) devices and machine vision systems that provide shop floor data. The limitations of the thesis as well as scope for future work has also been mentioned.

Sammanfattning

Smart tillverkning är ett koncept som har tagit över tillverkningssektorn. Definitionen av smart tillverkning eller det tyska konceptet Industri 4.0 har inte klart framgått och övergången från Industri 3.0 till 4.0 diskuteras fortfarande. Stora företag har börjat investera mycket pengar för att undersöka koncept som Industrial Internet of Things (IIoT), avancerad dataanalys och cloud computing, som komplement till smart tillverkning för att passa deras tillverkning och lösningarna är därför inte standardiserade. De har ändå lyckats förbättra sina produktionssystem för att göra dem mer flexibla, produktiva, effektiva och kundcentrerade genom att använda smart tillverkning. För små och medelstora finns utmaningar i form av låga investeringar, otillräcklig infrastruktur, ineffektiva produktionslinjer och begränsad tid för forskning om ny teknik. Denna avhandling studerar begreppet Industry 4.0 eller smart tillverkning och försöker skapa en övergång från Industry 3.0- till 4.0-konceptet med enkla lösningar. Huvudsyftet med avhandlingen är att skapa en affärsplan för att kommersialisera de enkla lösningarna så att små och medelstora företag på ett enkelt sätt kan börja införa smarta tillverkningskoncept. Två koncept, nämligen bildbehandlingsteknik och Bluetooth Low Energy-teknik studerades och en produktstruktur utvecklades. Användningsfall utvecklades och testades med produkten. Slutsatser drogs om hur produkterna och tjänsterna skulle gynna kunden och vilka problem den kan lösa på ett fabriksgolv.

En strukturerad affärsplan för att kommersialisera produkten med ett kundcentrerat synsätt utvecklades för att skapa en övergång från Industri 3.0 till 4.0 samt erbjuda kostnadseffektiva och enkla lösningar till små och medelstora företag. Valideringen gjordes med hjälp av Bluetooth Low

(4)

Energy (BLE)-enheter och bildbehandlingssystem för insamling av data från fabriksgolvet.

Avgränsningarna samt möjligheter till framtida arbete har också behandlats.

(5)

Acknowledgement

We would first like to thank Sture Wikman and Thomas Lundholm, the people who gave us an amazing thesis topic and guided us throughout the process. We were always given constructive feedback and it helped us make good progress and learn a lot. We also like to thank Per Johansson and Lasse Wingård for recommending us for the thesis.

Next, we want to thank the teachers and researchers at KTH for helping us with any queries that we had even though they were doing time crunching tasks of their own. We also thank the company representatives whose interviews were fundamental to our thesis work.

Finally, we thank our parents and friends for being extremely supportive throughout our years of study, providing us with unfailing support and continuous encouragement. This accomplishment would not have been possible without them.

(6)

List of acronyms and abbreviations

5G fifth generation wireless systems AI artificial intelligence

AR augmented reality

ASME American Society of Mechanical Engineers ASTM American Society for Testing and Materials BLE Bluetooth Low Energy

BM business model CAD computer aided design CCD charged couple device CNC computer numerical control CPS cyber physical systems DDS data distribution service ERP enterprise resource planning GE General Electric

GPS global positioning system HTTP hypertext transfer protocol ID identification

IEC International Electrotechnical Commission IFTTT if this then that

IIoT industrial internet of things IoT internet of things

ISO International Organization for Standardization IT information technology

KPI key performance indicators KTH Kungliga Tekniska högskolan

(9)

LTE long term evolution M2M machine to machine

MCS machine communication systems MES manufacturing execution systems MQTT message queuing telemetry transport MTC machine-type communications NFC near-field communication

OPC UA open platform communication-unified architecture OS operating system

PLC programmable logic controller RFID radio-frequency identification ROI return of investment

SME small and medium-sized enterprises SMS short message service

WiFi wireless fidelity WIP work in progress

XML extensible markup language

(10)

List of figures

Figure 1: Flowchart of the research methodology ... 12

Figure 2: Literature study ... 13

Figure 3: Validation of research questions ... 14

Figure 4: Beacons and gateway used [36] ... 23

Figure 5: Estimated number of beacons in 2020 [37] ... 23

Figure 6: Dashboard for beacon trigger settings ... 25

Figure 7: Example of an IFTTT application ... 26

Figure 8: Time stamps on an excel sheet indicating proximity detection ... 26

Figure 9: Data returned by firing a Rest API on Postman ... 27

Figure 10: Data showing access level of employees ... 28

Figure 11: Data showing the beacons detected at a location within a time period ... 28

Figure 12: Telemetry data from a beacon ... 29

Figure 13: Implementation of machine vision system ... 30

Figure 14: Notification from the machine vision system ... 31

Figure 15: Importance of Industry 4.0 for competitiveness now and in five years [39] ... 33

Figure 16: Graph showing the competitive analysis of the company with three similar companies [41] ... 34

Figure 17: Data for the calculation of ROI ... 35

Figure 18: Calculation of ROI ... 35

Figure 19: Graph showing expenses, revenue generated and gross profit for company ... 35

Figure 20: Industry 4.0 investment based on sector [39] ... 49

Figure 21: Annual investments in Industry 4.0 until 2020 [39] ... 49

Figure 22: Importance of Industry 4.0 for competitiveness now and in five years [39] ... 49

Figure 23: Technology investments based on activities/processes for digitization [39] ... 50

Figure 24: Graph showing the competitive analysis of the company with three similar companies [41] ... 52

Figure 25: Business model explaining problems and solutions solved by the product and services offered [41] . 53 Figure 26: Business model including key features for a business to function [41] ... 53

Figure 27: Data for the calculation of ROI ... 56

Figure 28: Calculation of ROI ... 56

Figure 29: Graph showing expenses, revenue generated and gross profit for company ... 56

Figure 30: Set-up costs calculation [43] ... 57

Figure 31: Profit and loss forecast ... 57

Figure 32: Cash flow forecast ... 58

Figure 33: Break-even analysis ... 58

List of tables

Table 1: Pros and cons of MTConnect ... 20

Table 2: Pros and cons of OPC UA ... 20

Table 3: MQTT v/s DDS ... 21

Table 4: Table comparing different companies working in the field of smart manufacturing ... 51

(11)

Introduction

Smart manufacturing is a concept that has become trending recently in the field of production engineering. Companies have been investing a lot of money to create systems that could improve their productivity, becoming flexible and customer-centric and in the long run, boost their profits.

Smart manufacturing deals with the use of data analytics, cloud computing and cyber-physical systems to improve the efficiency of a production system. Many large companies have used these concepts with favorable results, improving the productivity, reducing overall costs and created efficient production systems. This thesis deals with the research of smart manufacturing and looks at the concepts that compliments it. The aim of the thesis is to create a business plan for a company by investigating on technologies that simplifies these concepts in such a way that small and medium-sized enterprises in the field of manufacturing could start implementing digital strategies to improve their efficiency as well as productivity. The concepts used are simple solutions that could easily fit into any production system and it gives an initial threshold to further develop the concepts in a more customized way. Technologies such as machine vision systems and Bluetooth Low Energy (BLE) were mainly studied and considered as the base for the development of the product and services.

Problem definition 1: In the advent of smart manufacturing, there has not been a proper definition that has been formulated to explain the buzzword that is smart manufacturing or the German industrial synonym: Industry 4.0. This also means that there has not been a standard that has been developed to create a single system that could be incorporated in any factory setting to implement these concepts. Additional complimentary buzzwords like: internet of things (IoT), predictive maintenance and cyber-physical system (CPS) have also taken over the manufacturing industry. The transition from Industry 3.0, involving automation to Industry 4.0 involving data analytics, cyber physical system and cloud computing have not been clearly stated.

“How can small and medium-sized enterprises transition from Industry 3.0 to Industry 4.0 or smart manufacturing?”

Problem definition 2: The concepts of smart manufacturing, even at its infancy, have created an environment where companies who do not implement this new industrial revolution could in the long run be out of business as their productivity and efficiency could take a hit. The large industries have already started implementing or have started research in the field to improve themselves but the small and medium-sized enterprises who do not have the backing or the funding to improve themselves digitally could face problems if they do not adopt these strategies.

“How can small and medium-sized enterprises be provided with cost effective solutions to improve the efficiency, productivity and flexibility of the production systems by implementing the concepts of Industry 4.0 or smart manufacturing?”

The emergence of data analytics to keep track of everything has made companies to start mining data, which is considered to be the new oil of the modern era. In retrospect, most companies do not know what to do with that data and find themselves wasting money on something that creates no value to the system. The concepts discussed in the thesis uses data too but data that is simple and easy to understand for example: stoppages of a machine. With just knowing when the machine stops, it is easy to know how long the machine works and in addition it provides the efficiency of the machine. This data however simple is worth a lot to a smaller company that is

(12)

struggling to meet their daily production demands, and this can help to level out their production and improve their systems.

This thesis contains an extensive literature study of concepts that are pertinent to the field of smart manufacturing; methodology and case studies of different cases with initial concepts and a complete business plan for the setup of a company that deals with simplifying these concepts.

The products and services that the company provides are also tested out using a few scenarios.

The solutions are developed based on interviews conducted with three companies, who differ based on their size. The problems faced are considered as common problems that are faced by any manufacturing company and solutions were developed based on these problems. The focus is on producing simple retrofit solutions using new technologies with the aim of improving industrial technology solutions.

(13)

1. Methodology

Over the last few years a lot of advancements have occurred, and various technologies have been developed to make industries smart. For the development of machine monitoring, several important objectives were considered, and this section describes the aims and objectives of this research. It also includes a flowchart (Figure 1) of the research methodology, procedure followed for the literature study, questionnaire for the interviews, data collection and its analysis.

Problem definition

Data collection

• Literature review

• Interviews

• Trade fair and expo

Analysis of data and formulation of research question

Validation

• Asset tracking

• Employee tracking

• Telemetry

• Machine monitoring using machine vision system

Conclusions and future scope

Generation of business plan

Figure 1: Flowchart of the research methodology

(14)

Literature study

Before arriving to the methods that can be used to monitor different parameters of a machine tool, a systematic study of literature was performed. The focus of the literature study was to study about current advancement of the technologies that have been used and will be used in the future to make a manufacturing environment smart. Different sectors were studied from shop floor applications to IT systems, which can be inferred from Figure 2.

Figure 2: Literature study

Surveys and interviews

To answer the formulated problem definitions, a questionnaire (Appendix 1) was drafted, and interviews were taken, which can be found in the appendices. The main aim for conducting the interviews (Appendix 2) was to get data about how industries are dealing with the concept of smart manufacturing and if they have started implementing concepts of smart manufacturing to make their manufacturing facility smart. One small sized company, one medium-sized and one large company were chosen for the interviews.

Findings and validation

After conducting the literature study and interviews from different companies, various important research findings were established which can be seen in Figure 3. This led to the development of use cases for BLE technology and machine vision systems.

(15)

Analysis of data from literature study and

interviews

BLE technology Machine vision systems

Asset tracking Employee tracking Telemetry data

Monitoring of stack light using smartphone

camera Figure 3: Validation of research questions

Some research questions were formed which can be found in the Introduction and after analyzing data from the literature study and the interviews conducted, validation of the research questions were done by defining a few use cases which uses BLE technology and machine vision systems (3.3).

Generation of business plan

A business plan was generated to explore the future possibility when technologies for this thesis are commercialized focusing SMEs as the target customer.

(16)

2. Literature and state‐of‐the‐art study

The literature study considers eight different sections that contribute immensely to the foundation of this thesis work. Some of these topics are inter related and have a few sub sections that could be an initiation for prospects in the field of manufacturing.

2.1. Smart manufacturing and technology advancements

Smart manufacturing is about making products by managing the entire value chains along with product lifecycles, whereas the industrial internet helps in building, deploying and operating large connected systems. Smart manufacturing includes many aspects of manufacturing with full product lifecycles from the first idea to end of life [1]. The physical and digital world can be converged in the initial stages with sensors and sensory data, which automates and makes pattern tracking quantifiable for both product distribution and customer behaviors in the physical world.

These kinds of data are becoming the currency of the industrial internet economy [2]. As per the industrial internet of things (IIoT) vision of the world, a smart manufacturing enterprise consists of connected assets that work as a small part of a huge system or consists of smaller systems of systems that operate to some extent interactively and autonomously. IoT has different levels of intelligent functionality. It starts from sensing simple to controlling, optimizing and creates a full autonomous operation [3]. IIoT ensures flexibility, efficiency and profitability in manufacturing by closely linking smart connected machines with smart connected manufacturing assets with wider enterprise [3]. Asset performance can be improved by deployment of cost effective wireless sensors, easy cloud connectivity and data analytics. Smart manufacturing solutions will often require local processing of data at the edge versus central processing of data in the cloud due to safety, availability, security and data privacy concerns. Edge processing is generally provided at IoT gateway. In predictive maintenance applications, the latency requirement, and volume of data acquired may be prohibitive and edge analytics is the only viable option [4]. IoT, potentially could do far more than simply add efficiency to existing systems. New technology has already begun enabling entirely new business models. In addition to the classic manufacturing model characterized by design, product development, scale manufacturing, sales and service, there is the advent of “as a service” type products. Products are fitted with sensors and rather than selling the products outright and then separately handling responsibilities of warranty and maintenance, they are priced based on the hour. The sensors help to monitor the operating conditions and the time thus enabling preventive maintenance rather than fixed interval maintenance. This reduces downtime and the manufacturer can offer a simplified and more predictable cost of operation to its customers. Sensor technology, if used properly with actuators and robotics can enable preventive maintenance and self-healing systems, reducing costly downtime [5].

According to [6], the most common difficulty in building an analytical capability with respect to smart manufacturing is the lack of people with the expertise to conduct the analysis. Smart manufacturing provides mountains of data about customer demands and value chain logistics.

For the data to be useful, it must make sense and be used to boost the efficiency, grow closer to supply chain partners and develop products and services which the customers would buy. If this cannot be done, then the effort will be in waste. There are many challenges associated with smart manufacturing as it requires openness with data and collaboration, which in general would make some companies uncomfortable. Even though the concept has been present for some time, the technological and human skills required are often in short supply. The overall implementation requires a faith in large investments needed for products and processes involved where the end benefits that could be obtainable are unknown. General Electricals and Siemens are two companies that have started to solidify their position as platform providers for smart

(17)

manufacturing. They have developed cloud-based systems for connecting machines and devices from a range of companies, enabling transactions, operations and logistics.

The easiest level of managerial capacity that can be achieved with technologies and resources related to smart manufacturing is monitoring of production processes. Internet of things (IoT) can be used to map parts flow in real time by using strategically placed sensors at specific locations all along the production floor. Sensors and active RFIDs require expensive infrastructure which limits the extent of measured points, but it provides data during the entire production process and not just simply sampling collected data at different steps of production [7]. IoT can be used for remote assistance, for example: an expert remotely supports an operator via high definition, two-way AR video using high data rates and low latency [8].

From the perspective of operations, downtime has been considered the major problem for manufacturers. IoT offers a future with real-time visibility that would eventually eliminate machine downtime. Once, machines are connected to a single manufacturing platform, normal and abnormal states of the machine can be tracked more easily. A feedback system can be established with the manufacturing process using the data collected and using data analytics, patterns can be established. This will help improve efficiency, effectiveness and maintenance of processes [9].

5G refers to the fifth generation of wireless networks beyond the current 4G long term evolution (LTE) networks. 5G is expected to develop from and integrate with the older generations of wireless networks without any hassle. In first instance it represents an evolution of existing radio access technologies. The deployment of 5G networks is seen as important in fulfilling expected mobile data traffic growth. Improvements to existing mobile broadband services to provide enhanced mobile broadband (eMBB) is just one use-case for 5G [10].

The digitization of manufacturing fueled by CPS and the internet of things to enable connected and effective smart factories of the future. 5G could contribute to developments in the manufacturing sector by improving wireless data rates through the evolution of mobile broadband networks. It could take advantage of connection density and lower latency to enable multiple devices and machines to communicate with each other [10].

The potential use cases include automation by enabling devices to automatically communicate with each other enabling flexible and efficient production. Linking automated guided vehicles together improves better communication channels and improved safety. Using low maintenance sensors and actuators to wirelessly communicate with control units creates improved flexibility and efficiency with better inventory management. Logistics tracking of products from raw materials to finished goods can be made easily reducing cost and time. Robot control and remote assistance to fulfill operations such as measurement and digging can be easily incorporated with quick feedback. Augmented reality with live feed of a physical environment for maintenance and training can be done seamlessly using this technology [10].

The diversity of the manufacturing sector use cases provide opportunity as well as challenges.

Manufacturing companies are investing and partnering with Information and communications technology companies to drive innovation technology adoption in manufacturing. However, there are still questions as to who will provide the connectivity in controlled private spaces such as factories and how this might be charged for [10].

(18)

5G machine communication systems (MCS) enables the complete use of the IoT concept with the advancement of all machine-type communications (MTC), namely massive and mission critical [11]. MTC has a range of applications and services that include healthcare, logistics, manufacturing, process automation, energy and utilities [12]. These communications are enabled through a single infrastructure which is a prerequisite for mobile to mobile convergence communication protocol. Massive machine-type communication allows tens of billions of IP- based devices to connect through 5G MCS. This makes 5G an integral part of smart industry services and applications which include time-critical applications that require an immediate reaction. Industrial process automation and control is one of many use cases of mission critical IoT. Extra care in the design from the hardware, software, cyber physical system (CPS), to communication infrastructure and network architecture. The problem with the requirements of mission critical applications is that the state of the art IoT solutions evolved from the traditional embedded wireless sensor and actuator and the machine to machine communication protocol in LTE fall short [13].

Advancements in the field of automation has already improved processes and operations in various industries improving the efficiency. Automation can reduce human intervention drastically and thus speed up processes. These advancements are enhanced with the advent of mission critical communication which would create immediate reaction to crisis as well as create an environment for dynamic changes in an industry setting [14].

2.2. Cyber-physical systems

CPS are similar to IoT, which represents coordination between physical systems and computational elements integrated with internet. Implementing CPS in today’s factories offers several advantages that can be categorized in three stages: component, machine and production system. At the component stage, after the conversion of sensory data from critical components to information, a cyber twin will be responsible for capturing time machine records and synthesizing future steps to provide self-awareness and self-prediction. The next stage will have more advanced machine data like controller parameters, which are aggregated to the component’s information to monitor the status and generate cyber twin of each machine. In the production system stage, aggregated knowledge from the components and machine level information provides self-configurability and self-maintainability to the factory [15].

Sensors are the machine’s gateway to sense its physical environment and incase of sensor failure wrong or inaccurate information may pass to the algorithms which will give inaccurate results in the end. The developments of IoT and the emergence of sensing technology have created a unified information grid that tightly connects systems and humans together, which further populates a big data environment in the industry. With the combination of cloud computing and a CPS framework, future industries will be able to achieve a fleet-wide information system that helps machines to be self-aware and actively prevent potential performance issues [16].

2.3. Why predictive maintenance is important in smart manufacturing?

In this era of smart manufacturing, predictive maintenance along with artificial intelligence plays an important role. Predictive maintenance is not a new concept and is being used for many years without the knowledge of the term. A technician based on his experience, intuition and knowledge decides when to stop a machine and do repairs, which basically is predictive maintenance. In this increasingly digitalized world, every activity (internal or external) has some digital trace and data obtained from these activities and can be utilized to make sense out of it

(19)

with the help of artificial intelligence (AI). This application of machine learning algorithms for maintenance can be termed as predictive maintenance 4.0[17].

A study done by [17], indicate that 12% of the planned repairs can be avoided using predictive maintenance. It saves 30% of maintenance costs and 70% of unplanned outages.

The purpose of IoT devices is to get connected to every machine for data acquisition and this can be achieved using IoT platforms. Devices can be developed as per the need or be taken from common board developers like Arduino, Raspberry Pi etc. [18].

2.4. How work-in-progress is carried out in smart manufacturing

Radio frequency identification (RFID) materials can be used in manufacturing environment to monitor people, assets and equipment in the assembly line. It is simple to use and easy to implement for manufacturing companies to achieve real-time and consistent dual-way connectivity and interoperability between application systems at enterprise, shop-floor and work- cells. Devices equipped with RFID readers are active smart objects and those with RFID tags are passive smart objects. Smart objects interact with each other through wired and/or wireless connections, creating what is called an intelligent ambience [19]. Flow and traceability of information can be done effectively on a shop floor using RFID technology. The flexibility of the shop floors is obtained with little or no change in the layout [20]. RFID can be coupled with laser scanning to integrate shop floor motions as well as monitoring of parts moving through the production line. It will provide a 3D as well as a 360-degree view of the shop floor [21].

A line stoppage can be detected in the system using a context aware operator monitor app. It notifies the operator on the reason for the stoppage on the specific line event. The data source for the app is the raw historical data that is present in the system. The system for the stoppage notes down the time and duration of the stop, the sensor that caused the stoppage and a description of the resolution if handled solely. In another case, the field operator works with the line operator to resolve the issue. A mobile device fitted with a near-field communication (NFC) Reader is carried by the operator. As the field operator approaches the machine that is fitted with an NFC tag, the operator monitor app opens and the context aware system provides event data and information on possible causes. If the issue can be solved by the field operator, then the operator communicates back to the system which records the maintenance event. Another case is when a maintenance personnel with a maintenance app is involved in the process, where the maintenance app scans the NFC tag to retrieve information from the machine [22].

2.5. Monitoring using sensors and evolution of BLE

Temperature sensors are optimally used to capture the health state of the machine, enable predictive maintenance by detecting faults and use the data to anticipate future faults that might occur. The sensor equipment used is wireless as well as battery-free (piezoelectric transducers).

Sensing capabilities are provided by piezoelectric transducers through acoustic velocity dependence with temperature or strain in addition to identification and sensor signature with respect to a background clutter. Surface acoustic wave (SAW) sensors are made of a passive transducer probed through a far-field radiofrequency link, while data acquisition is performed by an associated interrogation unit [23]. The sensors are divided into five main categories: surface texture, surface integrity, dimensional accuracy, tool condition, and chatter detection. The sensors can be either in-process or in-cycle. An in-process sensor monitors the system during the machining process, whereas an in-cycle sensor examines the attributes periodically, such as

(20)

between parts. Mostly, indirect sensing systems are in-process while direct sensing systems are in-cycle [24].

Wireless sensors have transformed into BLE transmitters known as beacons, which have very high battery life and good accuracy. They consist of a combination of sensors that continuously monitor the process and transmit data into a system. Beacons are stationary devices whose signal is captured by the mobile whereas tags are Bluetooth powered tags that move around with assets or people and are located through stationary readers, scanners or gateways. Bluetooth 5.0 has slowly started being implemented but there are still problems due to the cost of implementing chips that can send and receive such powerful signals. Bluetooth in many cases is the next level technology for kanban systems [25].

Passive beacons are becoming more useful due to the high penetration of Bluetooth readers among the general population. Replacing dedicated and expensive readers with machine vision systems which are used by everyone will help to reduce cost and use better technology in an efficient way. Logistics companies should integrate beacons into their backend as well as front end systems. They must also link beacons to existing manufacturing systems, sourcing systems and other platforms in the long run [26]. Bluetooth low energy beacons can be used along with machine vision systems to provide data to decision makers regarding the state of the machine and the manufacturing process can be improved using the acquired data [27].

2.6. Machine vision system used in smart factory

Machine vision is a concept in which information from an image is extracted using some technology and method. In simpler terms, it’s an image processing technique which make automatic scanning of objects possible in a set field view. Camera is just like a sensor in a machine vision concept as it just gathers required data. Machine vision system is the one which uses image detection patterns and draw some conclusions based on the gathered data.

In an IoT enabled smart factory, sensors not only gather data or process image, but they also identify patterns and learn from them to make future decision. This can be related to predictive maintenance where user will be notified automatically when there is a need for maintenance for a machine or line [28].

Smart cameras are used in industries for data collection and image processing. A big advantage of having a smart camera is that they are equipped with sensors and have some processing power to do the image processing and thus the conventional method of sending images to a computer and doing the image processing is not required. They have application in the manufacturing environment for inspection, robot guidance, surveillance and tracking. Smartphone camera and webcam are used to in the manufacturing environment to do image processing and surveillance.

Unlike smart cameras, webcams send data to a computer to perform intense image processing as they lack processing power. Smartphone cameras got some potential to do some image processing as it has processing power and it is getting better each day, but its application is quite limited in the manufacturing environment.

2.7. Manufacturing execution systems and the language they communicate in Production monitoring and data collection are the main functions of manufacturing execution systems (MES). MES combines separate data collection and basically links shop floor and office systems. Real time production monitoring systems enable the continuous acquisition of data from

(21)

the shop floor about efficiency, malfunctions and productivity. This leads to improved production capacity and cost efficiency which helps to achieve the desired goals [29] .

MTConnect

Protocol designed for shop floor equipment to communicate to one another and to other computer programs that process shop floor data.

Table 1: Pros and cons of MTConnect

Pros Cons

1. Open source software

2. Communicates in XML and HTTP real time internet format

3. Plug-and-play interconnectivity between devices, equipment, &

systems

4. MTConnect monitors real-time performance of equipment and uses computational algorithms to detect opportunities for eliminating unplanned downtime and reducing planned downtime [30].

5. Best suited for CNC controls or other equipment that has known capability [31].

6. Provides standard dictionary for data terms no matter what machine tool builder is involved [31].

1. Not an instant solution, but a tool which has a learning curve and requires some study [32].

2. MTConnect is read only (writing back to the machine is not possible)

3. MTConnect require adapter to collect data from the machine and formats it for MTConnect agent. MTConnect adapters is not available for many types of machine tool [31].

OPC UA

The open platform communication unified architecture (OPC UA) is an http-based protocol for industrial communication as the unified architecture which ensures that automation systems are compatible with one another.

Table 2: Pros and cons of OPC UA

Pros Cons

1. OPC UA is a read-write protocol 2. The range of data available is greater

for OPC UA as it is not bounded by a dictionary of data terms like MTConnect

1. OPC UA does not have standard dictionary like MTConnect and user must know address of the running state inside the PLC. This location varies as per the equipment also which makes it harder for the user to find the location and may have to process data to get the desired information. Which

(22)

ultimately makes harder for the software to know the state of the machine or equipment [31].

Interoperability between MTConnect and OPC UA

MTConnect is a popular standard for machine tools, robotics and other equipment in a manufacturing environment. The MTConnect OPC UA companion specification has been developed by MTConnect and OPC UA community to establish a uniform information model that can be utilized via the MTConnect standard and the OPC UA standard to facilitate interoperability.

Connectivity and interoperability increase when MTConnect data model is applied to OPC UA.

OPC UA exposes the MTConnect data which opens a gate to connect to various SCADA and other software application on shop floor and the cloud. MTConnect-OPC UA integration increases customization and makes it easy to create solutions which are machine specific without jeopardizing compatibility with other equipment [33].

MQTT v/s DDS

MQTT is an ISO standard (ISO/IEC PRF 20922) publish-subscribe based messaging protocol. It works on the Internet Protocol suite. It is designed for connections with remote locations where a small code footprint is required, or the network bandwidth is limited.

DDS is a machine to machine standard that aims to enable scalable, real-time, dependable, high performance and interoperable data exchanges using a publish-subscribe pattern. DDS was promoted for use in IoT.

Table 3: MQTT v/s DDS

MQTT DDS

1. Telemetry: Device to server, data center, back office, IT cloud 2. Centralized and server-based

analytics, business logic and integration

3. Used in RFID

4. Remote resource monitoring 5. Centralized [34]

1. Intelligent systems within and between devices, dedicated systems and real-time cloud 2. Analytics, biz-logic and integration

distributed, embedded at edge 3. Used in manufacturing

4. Decentralized

5. Requires upfront investment [34]

Kafka bus

Kafka is an open source stream processing software platform written in Scala and Java. It is a program that is developed to handle real-time data feeds at high throughput and at low latency.

Its storage layer is essentially a massively scalable publish/subscribe message queue architected as a distributed transaction log, making it highly valuable to process streaming data. Kafka was developed based on the limitations of existing enterprise messaging systems. Messages are published by a broker to a topic. A consumer can subscribe to different topics from different

(23)

brokers and get data from the brokers. Kafka is a multi-subscriber system and a single message may be consumed multiple times by different consumer applications. Kafka uses a pull-based consumption model to process huge volumes of data that allows an application to consume data at its own rate and recall it whenever needed. By focusing on log processing applications, Kafka achieves much higher throughput than conventional messaging systems. In the field of industrial IoT, Kafka provides a powerful way to implement databus architecture. The APIs, which are the Kafka streams provide first-class abstractions for dataflow-oriented stream processing making it a data-centric application design. Together, they allow the system architects and other key stakeholders to position their apps and infrastructure to enable the future growth in relation to usage of sensors and data analysis while maintaining a clean architecture [35].

2.8. Summary of interviews

Interviews were conducted with three companies (Appendix 2): the first company was a small metal fabrication plant located in Flen, the second was a medium-sized impact socket manufacturer and the third was a leading manufacturer of heavy vehicles.

The owner of the small company realized the importance of smart manufacturing but due to the daily demands and their concern to keep up with everyday activities, they have not had the time to research on the possibilities. Being a small company, they do not use ERP systems and their orders come in small batches. There are a lot of opportunities to implement asset tracking, machine utilization and personnel tracking in the production facility.

The medium sized company has also not implemented a lot on smart manufacturing but would like to do so in the long run. The production manager understands that the systems are underutilized, and they are interested in using monitoring systems. The production manager would like to solve problems related to utilization of the machines as they are mostly starved for materials. Smart manufacturing and machine monitoring have great potential in the company and implementation can be done in phases but only after solving basic issues of production planning and levelling of mix. They also have problems with inventory management as some of the products that are made are shelved for a long time.

The large-sized manufacturer has a lot of resources to research and implement the concepts of smart manufacturing. According to them, creating digital twins would provide a very clear idea about a lot of problems in the systems without causing drastic effects on the real-time applications on the shop floor. Scenarios can be tested out and ideas can be implemented to obtain quality results. The tests are in the advanced stage and shop-floor implementation has begun in small scale with a possibility of becoming completely smart in the immediate future.

The interviews gave an idea about the scenario that is predominant in the industry, large companies with their resources research about implementing the concepts of smart manufacturing to get the most out of their system whereas small and medium-sized companies are left out.

(24)

3. Technical implementation

Based on the literature study and the interviews taken (Appendix 2), the problems were narrowed down and two major concepts like BLE and machine vision systems are taken into consideration.

These two concepts will be the basis for the product provided to the customers. A few real-world cases are tested and analyzed to check the viability of the product.

3.1. Beacons

Beacons are BLE devices that are used in many different industries as a replacement to high powered devices that use Wi-Fi. Beacons were initially used to send out push notifications to nearby Bluetooth devices notifying them of any offers in retail stores. Beacons are currently used in a lot of other industries like healthcare, supply chain, recreational facilities, manufacturing to name a few. Their reliability is attributed to their long life, minimal power usage and ease of use.

Figure 4: Beacons and gateway used [36]

With the advancement in Bluetooth technology and development of BLE, solutions like ibeacon and eddystone proved to be better than RFID in most of the cases. An active RFID solution when compared to solutions using Bluetooth beacons are found to be costly and impractical in large- scale operations. As per reports, beacon technology is expected to reach 400 million devices in 2020 [37].

Figure 5: Estimated number of beacons in 2020 [37]

(25)

3.2. Beacons applications

Beacons basically collect data from the point of application (big data) and this data can be useful in predicting machine health using real time alerts in case of failure. There is a lot we can do with low Bluetooth energy devices and some of the applications analyzed are described in the following sections.

3.2.1. Proximity detection

Proximity detection in beacons mean that whenever a beacon is detected or lost by gateway it interacts with user via a web panel or app developed by the beacon manufacturer. This information is quite handy in as it can be applied to a lot of things in a manufacturing plant. One application could be tracking of the product in a manufacturing environment by attaching beacons to an in-process product so when it finishes an operation or one complete set of operations in a cell and moves to another cell, a notification is sent out to the operator.

Another application could be using beacon tags for the workers which will give information about the motion of the worker in the location and movements can be monitored and compared with the pre-defined standard.

3.2.2. Machine monitoring

Beacons can also be used with manufacturing systems to gather data regarding the machines.

One advantage of using beacons is that the beacons contain a combination of sensors like temperature, humidity, accelerometer and vibration. Using this combination, a large amount of data can be obtained from the machine, which could help initiate predictive maintenance.

Beacons can be set to notify the operator when there is a change in the normal value with respect to ambient temperature, humidity or vibration so that the operator knows that the machine would have a possibility to break down soon and hence maintenance can be done during the idle time of the machine. This helps reduce possible downtime and improves the productivity and efficiency of the whole production system.

There are a few drawbacks as the location of the placement of the beacons play a huge part in the accuracy of the data obtained and this requires expertise from highly skilled operators who have worked with the machine for a while. With a good programmer, a lot more can be extracted out of the beacons with extensive uses of the telemetry packets that is available in them. Triggers and actions can be set with respect to proximity movements as well as condition changes. The triggers could be set as a threshold value or even the movement of an operator entering the proximity of the beacon. The action could be filling an excel sheet, an SMS, an email and a lot more. The technology is relatively new and a lot more can be achieved with respect to the beacons in the coming years.

3.3. Use cases 3.3.1. Asset tracking Introduction

Asset tracking is basically tracking assets from location, A to B with respect to shop floor in case of a manufacturing environment. Real time information about the asset or assets is vital in industries like manufacturing, healthcare, logistics etc. Asset tracking in industries is done by technologies like RFID, barcodes, WIFI, GPS, NFC or Bluetooth depending on the application. For example, to easily get information about the material leaving factory barcodes could be the easiest and cheapest solution. GPS is better for the outdoor tracking of the asset while NFC is best

(26)

used to communicate between tag and an electronic device when held close. WIFI tags works on the same concept as Bluetooth technology but take lot of energy when compared to Bluetooth technology.

Why beacons?

Beacons work on Bluetooth technology which is quite power-efficient as it only broadcasts information. A lot can be done with Bluetooth technology and it is cost effective when it is compared to other technologies like RFID and WiFi. Beacons are basically tracking devices which are coupled with sensors like accelerometer, proximity sensor, temperature sensor and humidity sensor. So, application of beacon technology is limitless and better than a lot of technologies which are used for the similar application.

Example case

To track assets on the shop floor, a beacon can be assigned to an item that has to be tracked from screwdrivers to big items like forklifts. If the asset leaves a location, the gateway can send an alert to the concerned person. This is an extensive way of implementing 5S in a shop floor so that important assets do not go missing.

Another case we could consider is inventory tracking. In a company that relies on fast delivery to market and a make to stock environment, there would always be a problem to track the amount of inventory that is piled up. Beacons can be used to keep track of these, as a group or as individual products. If the products are characterized as pallets, a single beacon can be used to notify the manager when the pallet has moved. For individual products, the implementations would be a bit harder if bigger beacons are being used. Simple technologies like buttons could also be used to track the inward and outward movement of inventory from a location so that unwanted stock and inventory can be avoided.

Example case is tested by using beacons and gateway manufactured by a Polish company called Kontakt. A scenario is imagined in which certain number of pallets are stored in a manufacturing plant with beacons placed on each pallet. Beacons are being monitored continuously by the gateway and a trigger is initiated in case of a missing pallet or if the pallet is at the wrong location to the concerned person.

Figure 6: Dashboard for beacon trigger settings

For this case, a trigger is installed to the beacons using Kontakt’s web panel and a mail response is generated by the IFTTT service. In case of beacon going out of the range from the assigned gateway, the authorized person gets an email stating that the pallet is missing from the location and a timestamp is recorded.

(27)

Figure 7: Example of an IFTTT application

A spreadsheet is also used for this case to get the different parameters like proximity of the beacon detected, unique ID of the gateway called as sourceId , beacon ID and timestamp.

Figure 8: Time stamps on an excel sheet indicating proximity detection

Conclusion

Beacon technology is still a work in development and the use cases have not been completely explored. Inventory tracking enables flexibility of manufacturing systems and helps to make the production systems more efficient and customize products to the customers in quick time.

Companies still rely on RFID scanners to track inventory and sometimes it is not effective. There are also a lot of disadvantages regarding beacon technology, but the advantages make it a promising and evolving technology with lots of monetary benefits to be gained when compared to RFID trackers.

3.3.2. Personnel tracking Introduction

One of the problems that persists in factories and shop floor is unwanted movement of people.

There is no real-time tracking of personnel inside or if they have entered a location that is restricted or if they must move a long way to get their required tools meaning more transportation which does not add any value to the products. Companies have a difficult time to follow lean principles when there is no value being added to the products and the whole system is quite inefficient.

(28)

Why beacons?

Beacons can be made in the shape of identity cards and given entry access to specified locations.

Employees with the ID cards can be tracked using gateways with very high accuracy. It increases security by automating and restricting unwanted access. It would give an idea to managers and supervisors as to how the employees move around and they can optimize their floor based on the data that is obtained.

Example case

Consider a case where the assets required by the operators, the machines for operation, the inventory are all in different locations and it requires the operator to move around the shop floor inefficiently, the beacons provide a simple solution to track how the operator moves around.

Based on the movement of the operator, the shop floor could be reorganized to minimize the movement of the operator and make the system more efficient. It is a modern way of implementing 5S on a large scale and thus implementing lean manufacturing in an effective way.

A scenario is created using beacons in which all the employees have beacons with different authorization access and IDs. First, a venue is created with access level 4 so employee only with access level 4 or higher can enter this venue.

Figure 9: Data returned by firing a Rest API on Postman

Using GET command to get data from REST API for the timestamp. Data is received with employees’ ID and authorization level which can be seen in Figure 10. A trigger can also be placed here so that when a person with access level lower than 4 enters, a mail would be sent to the higher authority.

(29)

Figure 10: Data showing access level of employees

Figure 11: Data showing the beacons detected at a location within a time period

Conclusion

Beacon cards are generally a lot expensive than normal ID cards and the technology being relatively new has a lot of flaws in it. The technology itself is quite promising but being expensive, not a lot of companies want to implement it in their factories or offices. The battery level of the beacons also poses a threat as they have a battery life of 14 months on an average. They could be used as a temporary solution to optimize the shop floor layout but as a security measure, the technology still seems a lot farfetched. If in the coming years, the price of beacons come down, it does seem extremely promising and a lot more advancements could be made on it.

3.3.3. Telemetry Introduction

Predictive maintenance is a topic of huge importance in the production industry in recent time.

This is mainly due to the loss of productivity of machines due to downtime and to prevent this from happening, companies are investing a lot of money on sensor technology to find the exact parameters that cause problems with the machine. The problem with this is, the complications involved with using sensors and there is a lot of research going on regarding the best location for

(30)

each sensor. Machine tool manufacturers have created advanced technologies that are making progress day by day to improve productivity and reduced downtime.

Why beacons?

Beacons consist of a combination of sensors in them and although not extremely accurate, they are still precise enough to work as simple solutions to find problems in machines with respect to vibrations or temperature. The solution is simple enough for small and medium sized companies who do not have the research backing or the money to spend on that kind of technology are often plagued by downtime and productivity issues. Beacons provide simple enough solutions to do the job or initiate a predictive maintenance environment in the shop floor and the technology holds a lot of potential if used in combination with sensor technology in the long run.

Example case

If a machine tool is working for a long time, the tool inserts are bound to wear off in due course, a beacon could be strategically placed somewhere in the machine to send a trigger to the operator when there is a deviation in the optimal value of vibration in the system. Readings from the accelerometer sensor in a beacon is tested and can be referred from Figure 12. Some triggers can be defined if the sensed vibration is more than the decided threshold. Beacons are also used to monitor proper working of robots as they would also start vibrating if there is some problem with the bearings in the robots and beacon technology could be used to detect these problems even before it happens, and the bearing could be replaced during the idle time and thus preventing downtime during production.

Figure 12: Telemetry data from a beacon

Beacons can also be used in a cell type production system, with each cell containing a beacon linked to the machine data like operating time, efficiency, temperature, speed and other parameters, which can be seen in the Figure 12. Whenever an operator comes in proximity to the machine, machine data is sent to the phones or tablets that are held by the operator from the beacons using BLE technology. This creates an interactive environment where machine data regarding a cell can be obtained effectively and efficiently with minimal power consumption.

(31)

Conclusion

The advancements of the technology are quite limited, and sensors still dominate the market. But for small and medium sized companies who have problems regarding downtime, this technology would be a huge step to reduce downtime of machines just by knowing when a tool could break or when a machine could stop working based on just the change in vibrations that the machine produces. The technology still requires a lot of thought to implement but it is still cost effective than overhauling the whole production line with new improved and more expensive machines.

This technology would be of great use to companies that use old machines and want to use simple solutions to be more efficient.

3.4. Machine vision

Machine vision systems are generally used to automate a machining process. It typically uses charged couple device (CCD) cameras and a vision processor which automatically detects process parameters that is fed into it or used as reference. Machine vision and industrial technology have revolutionized the way factories work and have provided solutions that solve automation issues as well as help implement predictive maintenance to an extent. Companies that are pioneers in the field of machine vision systems have dedicated solutions for any issues their customers come across but at the same time are extremely expensive to setup and maintain. This is where a simple solution such as a smartphone camera comes in. Smartphones are devices that are found in abundance and are usually underutilized based on their potential. Instead of expensive cameras, a few industrial applications were tested, where smartphones can be implemented to an extent, providing cost effective simple solutions.

Camera applications

The testing is done using a smartphone camera (Figure 13) with an app inbuilt to detect the change in color of light in the stack light of a CNC machine.

Figure 13: Implementation of machine vision system

The smartphone camera is calibrated first by placing the phone in a magnetic holder facing the stack light. The app is opened, and the lights are labelled and marked with a selection box. The

(32)

app works on the principle of image thresholding. Whenever the frequency of threshold is above a certain limit, a message is sent to a database (Figure 14).

Figure 14: Notification from the machine vision system

Stoppage detection

A stoppage is indicated by the machine, when the yellow or red light lights up on the stack light.

In case of the yellow light, the machine is waiting for the raw material to come in and whereas a red light indicates a failure stoppage of the machine. The transition of light is detected by the app and a notification is sent to the database. This notification can be sent as a message to the operator in charge. This makes up for the initial phase of the product that would be provided to the customers for testing out. The utilization and availability of the machine can be found out with the stoppage detection as well as change in the color of light in the stack lights which enables the detection of bottlenecks and helps optimize the production systems by rerouting parts to less utilized machines and reduces unwanted transportation, improving the productivity of the manufacturing systems.

KPI evaluation

According to [38], there are standard KPIs that are used for manufacturing operations. A few KPIs are chosen as the important parameters the company must monitor. Using the smartphone app, evaluation of important industrial KPIs could be done. The stoppages and the start times would give an idea about the actual production time and the actual unit busy time. This can be used to calculate utilization of the machine. Similarly, availability of the machine can also be calculated knowing the actual production time and planned busy time. These are just a few example KPIs that could be found out with just the stop, start and waiting times of the machine. In the advent of data analytics, where companies do not know what to do with the vast amount of data, the app can help companies improve their productivity with simple data that is easy to understand and manage. The data from the app can also be sent to a spreadsheet that notes down the timestamps and calculates availability, utilization and efficiency of the machine and provide graphs on a dashboard that is held by the operator in charge. This would-be part of the second phase of the product that is developed for customer use.

(33)

Other applications

In addition to using the app to evaluate light change of the stack light of a CNC machine, the app can be used with any machine that has a stack light associated with it. This means the application can be extended to assembly stations, maintenance call stations and even broadcast studios. The app used in conjunction with the vision technology will be incorporated with beacon technology for more advanced application cases in the long run.

(34)

4. Executive summary of the business plan

Introduction

The company helps small and medium-sized manufacturing enterprises to improve their productivity, efficiency and flexibility by providing cost effective solutions. It helps companies initiate the concepts of smart manufacturing in a simple way to make the production systems more productive. The company ensures that the SMEs do not miss out on the next industrial revolution and give them a chance to evolve and become better in the field of manufacturing. The vision is to become a leading technology solution provider in the field of manufacturing.

Management team

The management team will include an entrepreneur who has experience in managing companies as well as expertise in the field of manufacturing. A senior researcher with unparalleled competence in the field of smart manufacturing will be heading the technical aspects of the company. Their combined experience and Industry connections will pave the way for a smooth launch of the product in a real-life environment. A technical team of two production engineers and programmers will work on the development of the product for entry into the market.

Product summary

The product and services provided will be an app that uses the machine vision system to check the transition of light in a stack light of a CNC machine. The app is designed in price points with each price point creating new value to the customer, starting with an alert indicating stoppages to providing a clear visual description of efficiency and productivity of the machines. The camera can be used in combination with beacon technology for asset tracking, inventory management as well as initiate a predictive maintenance environment in the factory. A later stage combination with ERP systems is also taken into consideration to provide a whole smart factory solution to any factory environment.

Market summary

The product will be sold to small and medium level enterprises who are interested in the smart factory concept but do not want to invest their time or resources as they have bigger problems to cope with. Larger companies who are looking at simpler solutions can also use this product to improve the factory systems.

Figure 15: Importance of Industry 4.0 for competitiveness now and in five years [39]

Simplifying smart manufacturing concepts

Simplifying smart

manufacturing concepts

Commercialization of smart manufacturing concepts to help small and medium-sized enterprises

BHARAT SHARMA

PRAVEEN NATARAJAN

Simplifying smart manufacturing concepts

Abstract

Sammanfattning

Acknowledgement

Table of contents

List of acronyms and abbreviations

List of figures

List of tables

Introduction

1. Methodology

2. Literature and state‐of‐the‐art study

3. Technical implementation

4. Executive summary of the business plan