SMART
MANUFACTURING
AND
METROLOGY
How can metrology enable smart manufacturing?
Authors:
Supervisors:
Eric Tell
Andreas Archenti
Alexander Ökvist
Bo Karlsson
Abstract
To create the possibilities needed for more precise simulations and calculations regarding manufacturing changes in the equipment and new technology has to be implemented. This work investigates possible solutions for the gathering of information in manufacturing companies. To get a wider understanding of the current situation in manufacturing we have also researched some possible solutions and applications that can be applied in manufacturing. The work consists of a literature study regarding the possible solutions and technologies of smart manufacturing complemented by a survey and a follow-up interview with scientist and employees’ at large corporations to get their view of the business today and possibilities for the future. The benefits from a successful implementation of metrology can help companies toward success in the transformation toward smart manufacturing. This report also investigates what is needed for implementing smart manufacturing and the transformation in manufacturing companies to get economic advantages with a technological adaption. It also covers the possible difficulties and problems that may occur when this implementation is performed.
Sammanfattning
Content
1 The development of the industry 1
1.1 Complexity in modern manufacturing 2
1.2 Smart manufacturing 3
1.2.1 Metrology and smart manufacturing 4
2 Purpose, research question and limitations 5
3 Method 6
3.1 Literature study 6
3.2 Survey 6
3.3 Interviews 7
4 Theory 8
4.1 Cyber-Physical Productions Systems and cloud-based manufacturing 8
4.2 Manufacturing prognosis 10 4.3 Smart sensors 11 5 Applications 14 6 Results 16 6.1 Survey 16 6.2 Interviews 21 6.2.1 Present situation 21 6.2.2 Metrology 21
6.2.3 Data collection and analysis 23
6.2.4 Implementation 24
6.2.5 Examples 26
6.2.6 Challenges and the future 26
1 The development of the industry
To grasp the current state of manufacturing some historical information about the field is essential. During the 18th century the first real change in production began. Before this time most goods were produced by independent craftsmen without collaboration except for the few orders that could be placed by the military or the nobles. These workers created for their local market, which often was limited by city borders. Nations often imposed limitations on the establishment of new producers on positions already occupied to restrict competition, therefore resulting in less competition in existing industries. In extension, they also made sure that none of the producers gained a monopoly. This all changed with the first industrial revolution that started in 1740 in England (Encyclopedia 2017).
The challenges of today's industries are rooting from the changes that have influenced both customers and companies during the last century. The customers want cheap wares of good quality that at the same time can fit their specific needs. We are entering what the Germans call the fourth industrial revolution, the digitalization off the production industry. There exist difficulties to adapt these new technologies for a business that has a clear paradigm and a way of work regarding incremental innovation compared to the more radical changes of digitalization. According to PwC, 33 % of the industrial companies today see themselves at an advanced level of digitalization. By the year of 2020 at least 72 % of all industrial companies want to reach this grade of digitalization. In today’s industries only half of the companies are collecting data and big data analytics as a base for their decisions. These numbers are expected to grow to 8 out of 10 of all companies will use big data for their decisions in the next five years. (R. Geissbaur 2016).
The benefits of becoming a digital industry may seem far off but from the nine researched industries the amount invested would come close to 900 b.n $ for the coming five years. This number may seem daunting but they expected an annual return on at least 490 b.n $.
In present day, manufacturers are following the large amount of IT that is becoming an integrated part of production and measurement. The previous challenges to only maximize output are moving towards an interest in producing the greatest quantities of wares with the least amount of raw materials. Problems such as constant machine uptime, zero vision of defects and customizable products to specific need are all parts of the future challenges in manufacturing. Following these requirements, the problems of the future becomes clearer.
1.1 Complexity in modern manufacturing
Henry Ford’s breakthrough with zero complexity production has ended. Since the production of the T-Ford many manufacturers have competed by reducing complexity in their production, which have been successful. But those methods are insufficient to deal with the challenges of the future (ElMaraghy et al. 2012). Global competition and higher demands from customers have driven up the product and manufacturing complexity. The challenges today are characterized by design complexity that must be matched with flexible and complex manufacturing systems.
Complexity is driven by customer demands and expectations as well as global competition. The customers’ expectations of new and better products as well as services compels manufacturers to develop advanced products that is harder to produce and has higher demands on tolerances. To manage this the production methods will get more complex. In addition, global competition drives down the prices. Therefore, the manufacturing process must be cheaper, faster and strive to produce zero defect products.
Manufacturing companies operates in an uncertain and changing environment driven by changes in customer demands, product design and processing technologies. This increases the complexity in the manufacturing systems and is one of the main challenges for future production. The worlds private and public sector leaders believe that a rapid escalation of complexity is the biggest challenge confronting them. Their enterprises today are not ready to cope effectively with this complexity (Palmsiano 2010).
Due to the rising complexity in manufacturing there is a greater need of real-time data and knowledge of the processes. This data and knowledge can be used to anticipate and prevent problems in the process. According to Davis and Edgar smart manufacturing will lead to that production goes from response to prevention (Davis and Edgar 2008). On page 14 they wrote:
“…Response to Prevention addresses how sensors and knowledge-enabled capabilities will be organized and oriented. Every component of the enterprise will operate in a dynamic, proactive environment enabled by intelligent, model-based systems that are vigilant in monitoring plant and asset status. Any deviations from expected norms will be noted and if adverse trends are detected, the intelligent control systems will gather needed information and autonomously take preventive actions and in so doing exhibit high a high degree of fault tolerance.”
1.2 Smart manufacturing
National Institute of Standards and Technology NIST in United States of America defines smart manufacturing as:
“...fully integrated and collaborative manufacturing systems that responds in real time to meet
the changing demands and conditions in the factory, supply networks and customer needs.”
(NIST 2014)
The field of smart manufacturing consist of multiple different branches that all come together to form a new basis for production. It takes on the aspects in collection of big data, industrial connectivity, and the use of advanced robotics to further increase production and competitivity. The possibility to connect a whole factory and instantaneous transmission of data from one machine to the rest of the production. The usage of this data could also be applied beyond the factory itself and send data to either suppliers or customers to further integrate the process from raw material to finalized product (Jung et al. 2015).
Kang et al have defined some key technologies that are needed to implement smart manufacturing, they are cyber-physical systems (CPS), cloud manufacturing (CM), big data analytics, internet of things (IoT) and smart sensors. They also mentioned that these technologies affect each other in the implementation and usage of smart manufacturing (Kang et al. 2016).
IoT is one enabler for improving manufacturing, it can enhance automation, supply chain management and remote maintenance. It has potential to automate processes by connecting systems with machines, processes, and humans. It can also give direct access to design and manufacturing related data and information (Wu et al. 2015).
1.2.1 Metrology and smart manufacturing
Metrology is the science of measurement and its applications (BIPM 2008), the area has three sub-fields, scientific or fundamental metrology, applied, technical or industrial metrology and legal metrology. Figure 1 and Figure 2 presents how smart manufacturing sees metrology respectively how metrology sees smart manufacturing.
Figure 1:Smart manufacturing from the perspective of metrology (Szipka and Archenti 2017).
Figure 2: Metrology from the perspective of smart manufacturing (Szipka and Archenti 2017)
2 Purpose, research question and limitations
To know what to measure is one of the challenges of today's competitive industries. When these industries implement smarter systems and smart manufacturing this challenge might be the one that decides if the implementation will be successful or not. The factors that will have an impact on the production needs to be differentiated from measurements that will only waste money. If the companies do not know what to measure and how it can be measured they will have a hard time to stay competitive. To stay competitive companies will have to focus on what aspects they want to measure.
Measuring right parameters have been in focus since the implementation of quality control. In 1990 Guillot and Chryssolouris presented the advantages of using neural networks and the group method of data handling (GMDH) as tools when deciding what measurements to use in machine control (Chryssolouris and Guillot 1990). The purpose of this report is to identify some aspects of how data should be gathered and how this can enable smart manufacturing, the report also compiles a present day view on how smart manufacturing can be implemented by using metrology.
The main research question is:
• How can metrology support the development of smart manufacturing? Other relevant questions are:
• How smart manufacturing affects production?
• What systems will companies use and how will the systems collect and gain access to
relevant data?
• What data is relevant to improve the processes and make them more efficient?
3 Method
Data for this study were collected by reading articles, conducting a survey, and doing interviews with those who participated in the survey. The articles were both summarizations of current findings (secondary data) and articles with primary data. This literature study was done in the start of the project to get an overview of the subject and to find areas which could be studied further to answer the research question. Based on the data collected in the literature study the survey was designed and sent to three companies and two institutions. The questions in these interviews were based on the answers the interviewees responded with in the survey.
3.1 Literature study
The literature study was done to investigate the development in the smart manufacturing area during the 21st century and to identify what technology that could be used to help solve the research question. The data from the literature study were collected continuously during the research, in the beginning the literature focus was on summarizing articles and after hand more articles about specific subjects were studied. All sources are published articles from different scientific engineering journals except for two, which is white papers from a reputable consulting company and from a science foundation. Google scholar and KTHs database search system primo were used to find scientific publication databases, the articles were found by searching directly in a database or by searching on google scholar and then go to a database. The search words were: “smart manufacturing”, “Industry 4.0”, “Cyber-physical systems”, “CPS”, “Metrology”, “predictive manufacturing systems”, “predictive maintenance systems” and “industrial revolution”.
3.2 Survey
The goal with the survey was to get insight in what production companies, metrology companies and institutions view as important when developing a more efficient industry and how it can be done by combining metrology and smart manufacturing. Since the research area is not well explored the survey was designed as a qualitative study.
3.3 Interviews
4 Theory
4.1 Cyber-Physical
Productions
Systems
and
cloud-based
manufacturing
A cyber-physical production system (CPPS) is a system of collaborating computational entities which are in intensive connection with the surrounding physical world and its on-going processes. It provides and use data accessing and data processing simultaneously and can consist of several subsystems that are connected across all levels of production (Monostori et al. 2016). Three main characteristics of CPPS are intelligence, connectedness, and responsiveness. A CPPS needs to acquire information from its surroundings and act autonomously on that information. Therefore, it needs to be connected to sensors that gathers this data, but a CPPS must also have the ability to connect to other elements of the systems (including humans) for cooperation and collaboration. Monostori et al. makes the general assumption that a CPPS consists of two main functional components. One is responsible for the advanced connectivity that ensures real-time data acquisition from the physical world and the information feedback from cyber-space. The other one incorporates intelligent data management, analytics and computational capabilities that constructs the cyber-space.
Figure 3: The data flow in a CPPS (Lee et al. 2015)
Further, Gao et al identified some characteristics and benefits of CM, they are diversity, dynamic, virtualization, elastic, broad access, fault tolerance and cost effective. The three first that are mentioned can be key components to handle the new challenges with modern and complex manufacturing.
Both CPPS and CM are methods for data-driven decision-making and process control. In a conference paper researchers from NIST suggested that a control system can include actuators, sensors, and a controller (Helu et al. 2016). In Figure 4 the system is illustrated.
Figure 4: Industrial control systems (Helu et al. 2016).
The controlled process is disturbed and the sensors sends the data to the applications that analyzes the data and detects that the process must be corrected either by changing input parameters or by maintenance. The altering of input parameters can be done automatically or manually by the operator in the human machine interface. The controller receives the altered parameters and the alterations are realized by the actuators.
4.2 Manufacturing prognosis
Based on monitored conditions tool, machine or system lifetime can be estimated, this prognosis can provide a more scientific method to plan maintenance and create a more reliable production system design (Gao et al. 2015). For an example one efficient maintenance method, intelligent preventive maintenance (IPM), uses prognosis based on real time data to minimize down time and costs. IPM can increase system safety, extend machine life time, increase maintenance effectiveness, and reduce maintenance cost by preventing unnecessary replacements and machine failures that cause damage to surrounding machine components. Figure 5 describes how the architecture for CM can be designed. The machine conditions are monitored by sensors and sent to the cloud, the data from the conditions are analyzed by applications in the cloud or by a collaborative engineering team. The acquired results can form a base for preventive maintenance.
Figure 5: Architecture of cloud-enabled prognosis (Gao et al. 2015).
In their research, Gao et al mentions that studies have shown that preventive maintenance can reduce maintenance costs by 30 % and avoid up to 75 % of the breakdowns compared with using the regular norm with scheduled maintenance.
4.3 Smart sensors
A CPPS needs to collect data from the processes, that can be done with sensors or components equipped with sensors. One concept of smart sensors is gentelligent (GI) components, components that can collect, store, and communicate data (Denkena et al. 2010) (Denkena et al. 2014). The components contain manufacturing and quality data, stored as the “genetic information”. The intelligence of a genetic component is its ability to collect real-time data, process the data and then store and communicate this data. Denkena have divided real-time data into two types, configuration data and runtime data. Configuration data is generated from the engineering process to describe physical part and the machine. Runtime data is generated during the manufacturing process and describes the real-time status of the process.
The date that it stored as well as recorded in the components can be useful to plan and control the manufacturing, determine the cause of breakdowns, estimate when there is need of service based on the usage of components and aid in the planning to set up a dynamic service interval (Denkena et al. 2010). This possibility can reduce downtimes in manufacturing equipment and maximize the use of a component before it’s replaced. As illustrated in Figure 6 a closed information loop from design to maintenance can be obtained with gentelligent components.
In Figure 7 an illustration of how an gentelligent component can be manufactured. The production process can be more dynamic if the machines can read production data from the part they are producing. As written before can maintenance be more efficient with gentelligent components, for example can IPM be implemented.
Figure 7: Manufacturing planning and control with gentelligent components (Denkena et al. 2014).
Figure 8 illustrates how the control cycle of maintenance can be done with gentelligent components.
Figure 8: Control cycle of component status-driven maintenance (Denkena et al. 2014).
optimization of the process. Gentelligent machine tools and a monitoring system can make it easier to identify best practices for production sequences for every individual product in a batch. The best practice can change depending on the state of the whole production system, thus can gentelligent components make the production more flexible to change and disturbances. Denkena writes that an gentelligent driven production system can reduce lead times and interruptions in the manufacturing by 22 %.
5 Applications
Vogl (Vogl et al. 2016) tested a method for online monitoring of degradation in linear axes by using an inertial measurement unit (IMU). The method uses data from both accelerometers and rate gyroscopes to find changes and error motions in the axes. The method is said to be robust and appropriate analyzes based on the collected data can be done. The maintenance personnel can get information of the seriousness and location of the wear, further violations of performance tolerances can be estimated and an assessment if the axis needs to be replaced can be done. Vogl et al continued with suggesting that this method can be developed and supported self-diagnosing systems there among other useful information remaining useful life of axis can be estimated. In Figure 9 the concept of a mounted IMU and the data processing is illustrated.
Figure 9: IMU-based method for diagnostics of machine tool performance degradation (Vogl et al. 2016).
Okwudire et al (Okwudire et al. 2016) minimized vibration in a computer numerical control (CNC) machine by implementing a regular P/PI feedback controller augmented with velocity and acceleration data from accelerometers. This method can minimize tracking errors and thus increase process stability.
Aguade et al (Aguade et al. 2016) identified that direct measurement will be replaced by technology based on indirect measurement which requires shorter verification time. One of the measurement system suggested is laser tracker. In their article Aguade et al describes a method to improve a milling machines accuracy with indirect measurement. This can increase both process stability and product quality.
A tool condition monitoring system has four purposes according to Dimla and Dimla (Dimla and Dimla 2000), they are:
• “Advanced fault detection system for cutting and machine tool, • check and safeguard machining process stability,
• means by which machining tolerance is maintained on the workpiece to acceptable
limits by providing a compensatory mechanism for tool wear offset and
• machine tool damage avoidance system.”
considered the most applicable parameters. Dimla and Dimla also suggested to measure these four characteristics with optical methods, by measuring workpiece dimensions and surface qualities, spindle motor power consumption, magnetism, and finally ultrasonic methods. Dilma and Dilma’s suggested methods of monitoring tool wear in combination of a smart online system could automate the tool monitoring process, Ghosh et al (Ghosh et al. 2007) suggest a neural network to fulfill the same work. As seen in Figure 10 the artificial neural network (ANN) learn from manually obtained observation, after the system is trained the ANN can estimate tool wear in real time.
6 Results
6.1 Survey
The results from the survey are presented by identifying key words and present the number of occurrences of those keywords in the answers. If the answer to a question is accurate and describing the whole answer will be cited.
1. What do you think will be the main benefit of smart manufacturing in collaboration with metrology?
Citation: “Metrology plays a more important role in smart manufacturing compared with
2. To enable smart manufacturing, characteristics of products, systems, environment etc. needs to be captured and fed upstream. From this point of view which characteristics need to be captured by measurement?
Citation: “The same characteristics should be captured by measurement in smart
manufacturing as those in traditional manufacturing. However, the measurement needs to be made in a higher speed and also in a way of network, instead of individually, in smart manufacturing.”
4. How can smart manufacturing and metrology assist industrial practitioners in their work?
6. What can be the role of smartness in metrology to support application (in your relevant working field)?
Citation: “As a provider of qualitative and traceable data/information that are necessary for
the smartness of manufacturing.”
9. What problems do you see for companies that will not adapt smart manufacturing? Both now and in the future.
6.2 Interviews
Since the interviews had different questions based on the interviewees answers in the survey the answers are not directly comparable. Instead the interviews have been summarized and different subjects was identified, all that the interviewees said about the subjects is presented below.
6.2.1 Present situation
That improvement in today’s factories is possible was something that all the interviewees had in common with each other. Interviewee A identified the problems with manufacturing in present day was a reactive approach instead of working with the problems in a predictive and proactive way to hinder these occurrences before they even happened. Interviewee D spoke of the problems with both lead time and manufacturing time for products. These had to be improved in the factories to minimize cost and waste as well to maximize output.
Interviewee A and B had different situations regarding machine connectivity, interviewee A:s machines have been connected for some time and are transmitting some production data. Interviewee B on the other hand said that the earliest machine they could connect to the cloud was from the mid-2010s. Which in comparison is a newer machine park than interviewee A:s machine park. These differences might speak of possible complications with implementation in different kinds of manufacturing machines.
Interviewee B confirmed that there exists software to create digital machines and factories that can be fed production parameters. Interviewee A said that they already had digitized machines where simulation of production can be done. Both interviewee A and B said that the next step is to feed this digitized model with real time data from the production. Interviewee A explained how advanced simulation with production data can help the engineers introduce new products in the production. The simulation can tell the engineers how fast the process will be in the real world, but the hard part is to predict the capability since it depends on many factors. When a new product is produced it must be tested in a machine to evaluate its capability because the capability in the simulation is not good enough. E. g. it is not known if the processing of the part is efficient enough to have a high yield. If this could be determined before a test in the real world, resources and time could be saved and the ordinary production would not be disturbed unnecessarily. If production data could be gathered and use that data in the simulation program a better estimate of the capability can be done.
Interviewee A said that if in process measurement of the product dimensions is realized the production may become more efficient, if the dimensions can be corrected the machine will produce a correct product and if that's not the case any unnecessary machining is avoided. Regarding in process measurement of product dimensions interviewee D said that it is possible to some extent today, more complicated geometries will be hard to measure and it will be difficult to get the same accuracy as with a 3D-measurement machine. Interviewee A talked about one drawback of measuring the products dimensions directly in the machine. The machine compensate itself and if the compensation is wrong the machine might say that a product is correct even if it has an error. When the part is measured after the process in a separate 3D-measurement machine that specific error is avoided. But if the dimensions are incorrect, compensations must be done and a new part must be produced and measured to know if the process is correct. If the operator can be sure that the process is correct there is no need to measure the part after it is done, that would save money and increase the productivity.
6.2.3 Data collection and analysis
Interviewee C believes it is possible that all the parts of the user chain can benefit from sharing data. E.g. the machine suppliers get user data from the producer, the producers get user data from the end user. This can help both the machine producer and the manufacturer to improve their products to better suit the user.
Interviewee A suggested two different data collections strategies, one based on big data analysis and one based on selective data collection for specified problems. The strategies are presented in table 1.
Method Approach Advantages Disadvantages
Big data analytics
Collect everything that can be measured, analyze it and
then try to connect the results from the analysis to
what occurred during the process.
All the data will be available after the process
is done and it can be tied to every problem that
occurred during the process.
It demands a lot of resources and knowledge about the
process.
The operator and maintenance personnel will
not be included.
Selective data analytics
The engineers, operators and maintenance personnel defines X numbers of issues
with the process that they want to resolve, they also need to find out what caused
these issues and how to measure these phenomenon’s.
By doing this the X issues can be prevented.
Is not as resource demanding compared to
big data analytics.
The operator and the maintenance personnel will be included and they
might understand how this solution helps them in
their work. They could also provide their experience about the
You must know what you need to measure before you
6.2.4 Implementation
“When sensors are installed in a machine and the data is analyzed it will be like opening up the
box of Pandora.”
The citation is taken from interview A and the interviewee point was that when a company starts using sensors and analyzing data so many possibilities opens up and the company cannot be sure what they will be able to do.
Interviewee A believes that implementation of smart manufacturing will start with integrating simulation environments to the physical production, not necessarily the whole factory at once, one machine at the time is more sensible. The interviewee also identified maintenance as an easy implementation area with high yields. Interviewee B thought that implementation of smart manufacturing will start by installing machines that are already prepared with sensor for monitoring production process and are connectable to a cloud.
Interviewee A, B and C pressed the importance to connect the machines to a cloud. When the machines are connected all collected data can be transmitted to the cloud and be analyzed. Process quality can be evaluated and maintenance can be more dynamic. Interviewee C mentioned that not only machines in the same factory should be connect, different factories could also be connected to each other. Also, the operators and engineers should also have access to the same cloud as the machines. Whenever a decision is made it will be taken based on the information from the machines.
Interviewee A explained that not only the machine needs to be digital, human competence must also be digitized. When experience from an operator is put online will not only the analyzing systems and applications get smarter, an unlimited number of other operators and engineers will be able to use that experience to their own gains. This will lead to a more dynamic production according to interviewee A, the machine can analyze the process and if errors are detected the process can be stopped before the machine crash or ruins the product. Interviewee C believed that smart manufacturing will go through different phases, in the first phase will the operator will be engaged in the process. They will receive information from the machine interface and based on that they will take a decision. During the phases, the operator will serve the machine in the same extent as before but the machine will provide more and better information to the operator for every new phase. In the later phases the operator's competence is digitized and the machine can in the end take most of the decisions itself.
Interviewee C identified automobile and electronic industries as two of the main targets of smart manufacturing; there are already a lot of inspections and the industries are large scaled. The parts to the final products are created all over the world and it is important to keep track of the different batches. The implementation will not be easier but since there are already a lot of inspections in the processes of the automobile and electronic manufactures this will come more naturally. If those measurement instruments and sensors can be connected to the cloud and the data sent and analyzed, implementation of smart manufacturing is emerging. Therefore, interviewee C think it’s more motivated to invest in smart manufacturing for those two industries.
Interviewee B described the development of smart manufacturing as incremental innovations over time. To succeed with the implementation a clear goal and a strategy is needed. Interviewee C believed that it will depend on the industry. Because a car is built with several thousand part made in different countries, so if the factories all around the world are connected with each other you need a totally different way to solve this manufacturing connection. The problem is it can be at a to large scale even for these big companies. C thinks they have a very established production management system, but that can be totally changed with smart manufacturing, C thinks that for small scale manufacturing there can be incremental innovation but for those industries that has large scale manufacturing, there will be one big revolutionary innovation. Interviewee A concluded that it is time to start working with the different aspect of smart manufacturing. Even if a total implementation throughout the company will be expensive there are ways to harvest the low hanging fruit easily. Interviewee C thought that it can be difficult for companies to make decision when they should implement the different aspects of smart manufacturing. It is not necessary to adopt smart manufacturing early but it might be bad to adopt it late since the late adopters probably will lose competitiveness. There will be a balance of your investment and your gain. The timing is also important in this, C think that the sensors and measure instruments the beginning they can be very expensive so if companies invest early they may need to invest more and the cost will be reduced when the prices are lowered so those who adopt a bit later can invest lesser and get the same yield. That is the kind of difficult decision to make. It is a balance of investment and the benefits.
6.2.5 Examples
Interviewee A gave three different examples on how smart manufacturing can be implemented by using sensors and/or measurement:
Tool changes are executed about five to six million times per year in a manufacturing company. If the time it takes to change a tool is measured every time the operations is done any anomalies can be detected, when the operation takes slightly longer one or more parts in the machine might be worn and can be exchanged before they break. This is a very simple way to be smart, no extra sensors are needed so therefore it is a cheap way to measure the state of the machine. In the same way, a footprint can be taken every day, the machine does a certain routine and every step of the routine is clocked. The state of the machine can be checked without installing any additional sensors.
If the quality in the raw material can be measured the cutting forces in a milling machine can be adjusted to optimize tool life.
One example can be attaching sensors and a GPS to a forklift to be able to trace where the raw material is, every time the forklift places a batch of material a footprint is send to the cloud. Then some engineer realized that he/she could analyze the usage of all the forklifts on the plant, the result was that they had two redundant forklifts. The maintenance department also detected that every time the forklift drove into a pothole. Following this an order could be send to the maintenance department to fix the road.
Interviewee B and D gave one example each. B suggested that if spindle wreckage can be predicted by installing sensors for vibration and studying power supply to the spindle motor combined with empirical data any breakdowns can be stopped and the spindle can be exchanged before an incident. Interviewee D explained how production can be optimized by predicting the tool life left and when the tools condition is critical the machine can send a demand that the tool is to be replaced.
In total, three of five examples were about maintenance, and three of five examples was about process stability (one of interviewee A:s examples had both aspects).
6.2.6 Challenges and the future
Both interviewee A and B said that smart manufacturing will not be implemented in a standard package, it must be implemented differently for different companies because the different way of work that exists in the companies. Interviewee B added that it might be hard to motivate the investments needed to connect old equipment. Both interviewee B and C said that future machines need to be ready for smart manufacturing, the machines should already have advanced sensors and be able to connect to the cloud
6.2.7 Effects
All four of the interviewees said that smart manufacturing will make production more efficient. Interviewee A, B and D said one part of this efficiency will come from predictive maintenance and if the machine knows that the process was correct there will be no need to check the quality of the product after the process.
Interviewee A described the differences between older manufacturing streamlining and the streamlining from smart manufacturing. In the past process were streamlined by observing, listening as well as feeling and then some ad hoc solution were developed. The development with mereology and smart manufacturing leads to optimizing processes based on statistics derived from data. This development blurs the line between operator and engineer, the operator will do what the engineers have done until today. This will require the operators to have a higher competence than before. When you start to measure processes and machines you can see how they behave and you can make more accurate predictions about how the machine will work tomorrow and how the process will proceed in a near future.
Both interviewee A and B said that if the machine can evaluate the process and detect any errors in the product the process can be stopped before any more machining is done on it, this will make the production more efficient. Interviewee B also added that analysis of data can detect deformities in quality in real time but also track the problems that originates with a problem to better find out the reason behind. To gather all the data from the processes and upload this information to a cloud makes it possible to control the production in real time. To save the variables and parameters that a product was exposed to with collaboration with the process data will make it easier to draw conclusions about a possible wreckage. With this information, a company can withdraw specified product from a batch instead of the whole batch.
Interviewee A explained how investments can easier be motivated by using a simulation environment. With a simulation environment, a whole factory could be simulated and every change could be tested before it is implemented. Big investments would be easier to discuss and decisions would be easier to take if you know that the investment will work with the rest of the factory. This will also help managers see what investments that are important. The implementation process would also be shorter since all the tests already are done. With this digital model of the factory bottlenecks in the production can be illustrated and what steps in the production that are vulnerable to disturbances.
7 Discussion
The result from our research and survey are in line with what have been reported from other sources about smart manufacturing. The field is still in an early phase of implementation in various kinds of business. Actors that do not start to make a change now will most likely face the consequences of being a slow mover. Using metrology to enable this change towards smart manufacturing in an early state can be motivated by the economic benefits that will come from the various applications. The results from the survey are in line with the opportunities of the earlier presented applications. The industry needs better ways of predicting both their production and ways to be predictive with breakdowns and errors. These methods of measurement application rely on high standard of used measurement equipment that can be implemented in machines but we also have the conflict that the older machines in current production lack the opportunity to be fitted with said devices and connected to a cloud. From the interviews, we gained a lot of in depth information about the current state of production among leading companies. We were surprised about how far they had come in the applying the various solutions that we had been studying. The concepts that was theorized in recent scientific articles was already being investigated for implementation in current manufacturing and specifically CPPS had already seen some small scale implementation. The emergence of additional result regarding the profitability of these applications may further increase the speed of which companies will adapt the suggested applications to achieve smart manufacturing. The interviews also yielded examples regarding the problems with adapting workforce with the new technologies and applications. As the changes in workplace situations will probably become of a larger magnitude as the technological shift starts gaining momentum the investments from the companies to keep their employees educated and flexible will become important for a smooth transition.
So: How can metrology support the development of smart manufacturing?
8 Conclusion
9 Acknowledgment
10 References
Aguado, S., et al. (2016). "Improving a real milling machine accuracy through an indirect measurement of its geometric errors." Journal of Manufacturing Systems 40, Part 1: 26-36. Archenti, A. and K. Szipka (2017). Smart Manufacturing In the edge of Precision Engineering, Slide 5 and 6, recived 2017-02-11
BIPM (2008). International vocabulary of metrology – Basic and general concepts and associated terms (VIM). http://www.bipm.org/en/publications/guides/vim.html.
Chryssolouris, G. and M. Guillot (1990). "A Comparison of Statistical and AI Approaches to the Selection of Process Parameters in Intelligent Machining." Journal of Engineering for Industry 112(2): 122-131.
Davis, J. and T. Edgar (2008). "Smart Process Manufacturing An Operations & Technology Roadmap."
Denkena, B., et al. (2010). "Genetics and intelligence: new approaches in production engineering." Production Engineering 4(1): 65-73.
Denkena, B., et al. (2014). "Development and first applications of gentelligent components over their lifecycle." CIRP Journal of Manufacturing Science and Technology 7(2): 139-150. Dimla, S. and E. Dimla (2000). "Sensor signals for tool-wear monitoring in metal cutting operations—a review of methods." International Journal of Machine Tools and Manufacture 40(8): 1073-1098.
ElMaraghy, W., et al. (2012). "Complexity in engineering design and manufacturing." CIRP Annals - Manufacturing Technology 61(2): 793-814.
Encyclopedia, W. (2017). "Industrial Revolution." Europe, 1450 to 1789: Encyclopedia of the Early Modern World. Retrieved April 6, 2017.
Gao, R., et al. (2015). "Cloud-enabled prognosis for manufacturing." CIRP Annals - Manufacturing Technology 64(2): 749-772.
Ghosh, N., et al. (2007). "Estimation of tool wear during CNC milling using neural network-based sensor fusion." Mechanical Systems and Signal Processing 21(1): 466-479.
Helu, M., et al. (2016). "Enabling Smart Manufacturing Technologies for Decision-Making Support." (50084): V01BT02A035.
Jung, K., et al. (2015). "Mapping Strategic Goals and Operational Performance Metrics for Smart Manufacturing Systems." Procedia Computer Science 44: 184-193.
Kang, H. S., et al. (2016). "Smart manufacturing: Past research, present findings, and future directions." International Journal of Precision Engineering and Manufacturing-Green Technology 3(1): 111-128.
Lee, J., et al. (2015). "A Cyber-Physical Systems architecture for Industry 4.0-based manufacturing systems." Manufacturing Letters 3: 18-23.
Monostori, L., et al. (2016). "Cyber-physical systems in manufacturing." CIRP Annals - Manufacturing Technology 65(2): 621-641.
NIST (2014). "Smart Manufacturing Operations Planning and Control." Retrieved April 4, 2017.
Okwudire, C., et al. (2016). "A trajectory optimization method for improved tracking of motion commands using CNC machines that experience unwanted vibration." CIRP Annals - Manufacturing Technology 65(1): 373-376.
Palmsiano, S. J. (2010). Capitalizing on Complexity, Global Chief Executive Officer Study, The IBM Corporation.
R. Geissbaur, J. V., S. Schrauf (2016). Industry 4.0: Building the digital enterprise, PWC. Vogl, G. W., et al. (2016). "Diagnostics for geometric performance of machine tool linear axes." CIRP Annals - Manufacturing Technology 65(1): 377-380.
11 Appendix
Appendix 1 - Survey questions
1. What do you think will be the main benefit of smart manufacturing in collaboration with metrology?
2. To enable smart manufacturing, characteristics of products, systems, environment etc. needs to be captured and fead upstream. From this point of view which characteristics need to be captured by measurement?
3. Should these characteristics be measured before, after or during the process?
4. How can smart manufacturing and metrology assist industrial practitioners in their work?
5. In what areas do you see easy/easier implementation and high yields for metrology in manufacturing?
6. What can be the role of smartness in metrology to support application (in your relevant working field)?
7. What are the biggest obstacles in the concept of smart manufacturing where metrology can offer solutions?
8. Can you describe the timeline for today’s factories to integrate metrology to the smart manufacturing concept?
9. What problems do you see for companies that will not adapt smart manufacturing? Both now and in the future.
Appendix 2 - Questions Interview A
1. Vad kommer man kunna förutse?
2. Vad kommer man kunna jobba proaktivt med. 3. På vilket sätt blir processerna stabilare?
a. Vad gör Predictivity , proactivity och process stability till de främsta fördelarna med smart manufacturing?
4. Vad för information från maskinen behöver du veta? Hur mäter du det? a. Vilken information från verktygen behöver du? Hur mäter man det?
b. Product quality är brett, vad är det man ska mäta för att få information om kvaliteten på produkten? Hur mäter man det? Vad mäts i dagsläget?
5. Vad är det man analyserar?
a. Vilken statistik är hjälpsam/relevant?
b. Ser du några problem/motstånd med dessa typer av hjälpmedel? T. ex. Man måste byta arbetssätt/vanor och lära sig nya saker. Det innebär en omställning i arbetet
6. Varför tror du att smart manufacturing är lättast att nyttja inom underhåll? a. Finns det några andra områden som du tänker på?
7. Hur kan omställningen från analog till digital mätning se ut? 8. Vilka problem inom smart manufacturing kan metrologi lösa?
9. Kan du föreställa dig en tidslinje för implementationen av smart manufacturing? a. Ska man vara tidig med att implementera smart manufacturing eller ska man
vänta tills det är beprövat? Vilka skillnader är det för stora och små företag samt är det skillnad på om företaget är världsledande eller ej?
b. Är det tillverkande företag eller maskintillverkarna som styr tidsaspekten av implementationen av smart manufacturing?
10. Varför kommer dessa företag inte att vara lönsamma? 11. Hur kan detta locka aktörer till området?
Appendix 3 - Questions Interview B
1. Vilken fakta/data behöver vi veta för att ta ett beslut. Kan du beskriva en situation som är typisk för dig eller en stor aktör?
a. Hur föreslår du att man ska mäta vibrationer, mekanisk förslitning och b. geometriska parametrar?
c. Vilka produktionsparametrar är viktiga?
2. I vilket skede tror du det viktigast att mäta dessa egenskaper (products, systems, environment etc), under eller efter processen?
3. Har du några idér om hur en implementation i dessa maskiner kan gå till? (single and bottle-neck machines)
4. Vad finns det för stora hinder för att smart manufacturing kan implementeras och hur kan metrology lösa dessa hinder?
5. Om hur lång tid tror du att det kommer finnas maskiner som är gjorda för att kopplas upp mot IoT och samla in och analysera data?
Appendix 4 - Questions Interview C
1. How do you think this foundation can be build? Lots of companies wants smartness in their production but may bypass important technological foundations.
2. How can the change from traditional measurements to digitalized and connected measurement devices be applied in today’s factories?
3. Companies have problems adapting their current machines to suit the needs of smart manufacturing. In today’s environment they instead are phasing out the older machines for newer models that have better modifications to allow smart manufacturing.
4. Can you think of any possible ways to connect the older machines to allow a quicker change towards enabling smart manufacturing?
5. These aspects will change the way employees work in a producing company. What are your ideas to help the people working there to easily implement and adapt to the changes in their working environment?
6. Why do you think it will be easier for automobile and electronics industries to implement metrology in smart manufacturing?
7. What will results in higher yields? (Cost efficiency, low waste, high machine uptime etc.)
8. How do you think an integration of a whole supply chain would work if everyone could share the data? Eg. machine producers get usage data from the production company and the production company gets user data from suppliers and end users.
9. What are the benefits? 10. What are the disadvantages?
11. These measurement instruments and sensors for proper measurement in complicated manufacturing processes, are they available in today’s market or will production companies have to invest to fit their specific needs?
12. How can productions companies then attract the desired employees with the education to create these instruments/sensors?
13. Will the concept of smart manufacturing be a field of many incremental innovations or a few radical changes? Please motivate why.
14. Those industries that adopt smart manufacturing late or not at all, will they lose competitiveness due to higher productions costs?
Appendix 5 - Questions Interview D
1. What do you mean when you say through-put?
2. How is through-put one of the biggest issues for smart manufacturing? 3. How can this problem be solved?
4. Why do you think vibration and temperature drift will be important variables.
5. How will operators be assisted by smart manufacturing and metrology? Feel free to give one example.
6. How can metrology help to minimize stops in production?
7. Do you think that the concepts of VR/AR can help enable remote maintenance?
8. Do you think that the threat of security breaches stops companies from adopting smart manufacturing?
9. When do you think a machine (for example a milling or turning machine) will built to be fully connected to a cloud service, be able to measure all process parameters, feed that data to the cloud and receive feedback from the cloud to optimize the production process?
