Machine Learning in Equipment Maintenance

Brief overview the project

The project revisits the purpose of the project, repositions it in the course of the research problem and expounds through research aim and objectives. More attention has been given to the relationship between machine learning approach and reduction of the preventative maintenance of the equipment as far as monitoring is put into consideration. In the course of covering the scope of the research, the project has already highlighted some of the research tools that are felt relevant to the research process.

Core terms and concept of the research

The research paid attention to some of the terms and concepts that could be applied in the research process. Some of the concepts and terms covered include machine learning, artificial intelligence and predictive maintenance among others as described below.

Machine learning: The research understood machine learning as part of the applications of AI which enables systems to automatically learn as well as improve in the experience without necessarily being programmed [1]. The research took advantage of this understanding in exploring the advantages machine learning would provide to systems applied in predictive maintenance.

Predictive maintenance: It is a technique applied in predicting future failure to an extent of the machine component thereby forcing the replacement of the component [2]. This understanding has appeared in the research as part of the prevailing systems that target reduction of costs through maintenance than repairs which would be catered for in case of a breakdown.

Artificial Intelligence: This is also one of the concepts that have been applied in the research to denote simulation of the human intelligence by use of machines, which can be computer systems among others [3]. The meaning of this concept has helped in further understanding machine learning and changes that are likely to influence the outcome of the research.

Machine learning & industry maintenance: The research has already searched for materials that reflect on the essence of machine learning in the industrial context [4]. Through literature and methodological findings have helped in building an understanding of the real application that reflect case scenarios where machines learning have served or helped in handling an industrial problem.

Background:

The Background section of the project demands a closer look at the preliminary studies that would provide a glimpse of the idea behind the research topic, and research gaps that could be identified. Perhaps, the pre-visit of significant material is needed to gain the necessary knowledge at the beginning of the research process. Currently, the pre-visit has been achieved with key mention of machine learning and its application in predictive maintenance. The background study has also offered a consistent coordination of the case scenarios, which interlinks one application to the other. Such interlink is necessary for two basic reasons. First, it is necessary to have a flow of ideas with growing relevance to the research topic. Secondly, it is easier to establish gap areas from the case studies from literature such as transportation, telecommunication, and financial market. In this context, the two have been achieved with gap areas indicating a missing direct application of machine learning in predictive maintenance without involving other technologies.

Outlines and Justification of Research Problem

The research has already shared insights regarding the research gap areas in terms of direct application of machine learning. More attention has been given to case studies that noted the essence of other technologies that work in line with machine learning. Identification of such gap areas in research defined the direction for this research which aids the process of informing gaps across literature.

Importance of this research

This area taps into the gap areas noted in the background section and turns them into problems that need solution through research. A slight revisit of the background is convenient and puts the research in the right course. The context has already addressed this area by confirming that case studies have been relevant in sharing significant knowledge in the highlighted areas. Some of the highlighted areas include predictive maintenance, neural networks and artificial intelligence. The interaction of these highlighted areas confirms the essence of interconnected disciplines or areas of knowledge, which produce probable solutions to the anticipated research problem. The research proposal has already covered this milestone and went head touching on the area of focus.

Research Aim and Objectives

The section demands development of the research aim and the supporting objectives. Perhaps, the research aim is developed from the research problem. Objectives only provide a breakdown of the research aim into more simplified parts. The proposal has already developed this part while touching on the key concepts that are more relevant to the research process. Three supporting objectives have been aligned to the research aim while observing the key principles and words felt appropriate in the context. The objectives have been used to break down the research aim into researchable components.

Progress

Literature review takes into consideration the theoretical, empirical and methodological contributions made by other case studies related to the research topic. This prerequisite attracts insights on machine learning and industry 4.0 in the face of machine learning. A general facet of this overview attracts case study findings, contributions from research-based articles and reports. While this part could not tap into comparative studies, it still remains relevant in connecting the background study to relevant findings tapped from other researches. First look was into the application machine learning in different field then taking each as a case study on wider industries and applicability into predicting future based on collected data. Prominently, machine learning has had significant influence in businesses sector, financial market, healthcare, transportation (railway and airline), monitoring rotary machines, and communication industry (telecommunication). Through use of big data and artificial intelligence, machine learning can predicts the future of a given scenario such as performance of financial market. Such case studies from reviewed documentation and literature is highlighted through this report.

Machine learning, predictive maintenance, and AI

The research noted that when the theories, concepts and principles are searched in a wide range of materials, it is possible to think about the research problem in various ways. This helps to build a better research. Some of the concepts are not limited to predictive maintenance, machine learning as well as artificial intelligence among others. The coverage of these elements helps in understanding the essence of technology and the need to review the most updated materials [5]. While the concepts and terms appear in different context and would be used to imply different meanings, most of the working principles remained fixed and showcased a consistent trend in a range of the research materials. More attention given to machine learning and predictive maintenance has already revealed commonalities with one platform believed to have made use of the technology before [6]. However, it is hard to cover the relationship across the concepts under respective sections given the independent applicable use of each [7]. Despite the challenge, the coverage of each concept helped in understanding an informed background of information associated to the respective components of the research. Such an understanding helped in reconnecting the concepts, terms and theories while constructing the conceptual framework for the research process. Again, such an understanding helped in constructing simple relations that could further be researched and the materials reviewed to gain insights of the research aim and objectives.

Role of machine learning in predictive maintenance and anomalous detection

This is a specific part of the literature review that attracts researches and case studies such in telecommunication and financial market in a particular area. Apparently, this specific component is connected to the research aim, which touches on the wide area of the research topic. While reviewing the role of machine learning, the proposal has already searched for concepts, theories and case study findings aligned to predictive maintenance and the role of machine learning. The only remaining side is to integrate the findings and document them under the literature review section of the research.

Data gathering methodologies for machine learning in industries

This is also another section that attracted the attention of the proposal. First, the research has borrowed insights from the background study, which already shed some light on machine learning and the relevant industries. The proposal has also prepared several research documents that can address this research area. Among the material sampled are the peer reviewed articles and journal articles among others. The searching process considered the time a document was published, relevance of the case studies and the nature of the findings. The only missing part is to document the findings and the methodological contributions.

Influence of machine learning in sparing costs in the production and manufacturing process

Attention has been paid towards the impact machine learning can have in terms of sparing costs in the course of manufacturing and production process. The proposal has already assembled the necessary materials said to await the review process. However, it is worth noting that this is one area that attracted the least number of materials when compared to other sections. Nevertheless, the proposal still maintains the standards engaged in selecting the material. The only missing part is to document the theories, findings from case studies and concepts aligned to this section.

Ways through which machine learning can reduce maintenance activities

This is the last mile of the literature review that pays close attention to means that machine learning use in reducing the maintenance activities. The proposal has already made significant strides towards reviewing significant case studies attached to this section. First, the proposal has assessed the relevance of this section to other sections under the review. Secondly, the proposal has assessed the commonality of the materials searched before and the ones needed in this context. Besides, the proposal has added other materials to the common ones tapped from other sections. The missing side includes documentation of findings, theories and concepts.

Challenges faced in identifying and sampling out literature

In the presence of search engines and open source databases, it is easier to find materials linked to a particular subject topic. It is even simpler to search articles on such databases like Google Scholar among many others. However, there are still challenges one would encounter while sampling the literature. First, the research topic was slightly long and this could not attract significant findings. Perhaps, when the item to be searched is too long, then most of the databases would rarely capture the most relevant data. Then the really challenge comes in when establishing relationships among concepts. For instance, how to coordinate machine learning and predictive maintenance is more critical while making an attempt of tapping into relevant and significant materials. Using Boolean such as AND/OR may not obviously give the desirable results and this compels the researcher to look for extra words that would be added to the relationship before realizing the outcome. For instance, if the search phrase is “machine learning and predictive maintenance”, then the results would not be the same like that of “the influence of machine learning on predictive maintenance”. However, it is sometimes not easy to notice the differences thereby tapping into less substantial information. Lastly, the lately published materials are always recommendable in giving the true face of the current research. However, some of the old materials could still have significant information but may be ignored due to the presumption of outdated details that may ruin the course of the research. This may not be true in some of the cases especially when old materials sound more informative compared to the latest ones. It sometimes becomes hard to make such tough choices unless there is a guideline in searching for the material.

Summary (so far)

The above challenges could be handled through an established guideline that would help to move the research closer to the most relevant research findings. While the position of literature is to tap into some of the empirical studies, the project has already attained this by considering case studies published in the most recent years. The progress has attained most of the quality requirements including coverage of the theories and concepts in the recently published materials. More attention is given to the methodological and empirical findings which could easily be spotted across the sampled materials. In addition, the research turned the objectives into researchable areas, which needed to have a background which would easily be compared to the findings that would be realized at the end of the research process. This is one important area that has already been highlighted in an effort to produce justifiable findings

Development of research Methodology

Research philosophy

The research has already done a review of some of the research philosophies available for the research process. Some of the reviewed philosophies include positivism, interpretivism, realism, critical theory and the pragmatic paradigm [8] [9]. The assessment of this class of philosophies noted that positivism is inclined towards quantitative research while interpretivism is attached more qualitative research [10]. However, a further analysis of pragmatism confirmed that the paradigm is more relevant to the mixed methods research adopted in this context [11]. However, this choice could still be accompanied by highlighted areas of strength and weaknesses as well. Despite identifying areas of weakness, the determination of the research philosophy helped in determining the course of the research by establishing the ontology and epistemology before determining the typical methods that could be used.

Research methods

The choice of the research method has already been done but not yet documented. In essence, this is due to the progressive reports that have to show progress every time without running more than two tasks at the same time. However, after considering different cases that showed similar characteristics to this research, mixed methods were seen to be more appropriate in defining a collection of tools needed for the research process [12, 13]. With no revisions done to the selection of the research method, it is obvious that this is still open for recommendations.

Research approach

This section is also done in the sense that both deductive and inductive reasoning can be applied. This section has already been described in line with the research method and the philosophy adopted for the research process [12]. The description aligned to the selection has indicated the advantages of this choice and the limitations that need to be observed while handling the research.

Data collection

This section has already been reviewed and significant choices done with the remaining part entailing documenting the details and reasons behind them. Primary and secondary data research would be put to task. This is also accompanied by identification of the target population which awaits mapping and the relevant procedures that need to be performed before engaging the interviews. Details of this section shall equally be reflected in the report. Same attention has been given to the process of designing the secondary data research which equally plays the most fundamental role.

Ethical considerations

Given that the research will conduct primary research, chances are that the researcher has to interact with human participants. The research has therefore put into consideration the ethical guidelines that would be used before, during and after the research process [14, 15]. The research is already aware of the standards, measures and principles that need to be observed for the purposes of avoiding chances or risks of putting the participants’ in the harm’s way [14]. The participants’ information sheets are ready and even the consent forms have been finalized just awaiting the research to commence as scheduled.

Progress

The proposal treats this section as one of the most determining factor of the research process. However, only a glimpse of the research method, research design and tools for data collection has been indicated in the brief. It is worth noting that progress has been made in terms of developing the kind of questions that will be involved in the interview process. The inclusion exclusion criterion is on its final stages as far as secondary data research is put into consideration. The research has also paid attention to the need of seeking permission from relevant authorities. In addition, the research has designed the Participant Information Sheet and the Written Consent forms, which can be shared to the relevant participants before the eve of the research.

Timeline of the Activities So far

Breakdown of the activities completed and in-progress so far

The research proposal prepared a Gantt chart which has showcased the kind of activity, the starting date and the duration assigned to each one of them. Some of the activities include studying and ideation, proposal draft, proposal review, literature review, designing methodology, designing data collection tools, reviewing the research process, collection of data, data analysis and documentation. Some of the tasks that have already been completed include ideation, literature review and the review of the proposal. Activities in the pipeline include expanding the literature to explicitly cover the keynote areas and partially designing the research process. The Gantt chart has already shown the duration for each activity while ensuring that there is no bombardment of activities. However, this can still be revised based on the changes of the conditions on the ground. The activities in progress include the extensive literature that intends to cover every substantive area that may have an impact on the research process. Next activities include designing the methodology, which is important in defining the tools that are important in establishing the research process.

Challenges encountered and steps taken

While the process seemingly stood out more successful, there are still challenges that could be encountered during the process. First, in the course of conducting the literature review, it could be noted that establishing relations among the concepts was a bit challenging. Again, the choice of dumping the outdated materials was challenging given the rich content in some of them. However, the research chose to establish a guideline that would engage the materials and tap the most significant information and findings from them. Again, making choices of methods to be used in the research process posited a dilemma with most of the tools sharing common characteristics. This could however be determined by first, studying the paradigms which carry with them the ontology, epistemology and typical research methods that could be applied along.

References

1] Harrington, P., 2012. Machine learning in action. Manning Publications Co..

2] Lee, W.J., Wu, H., Yun, H., Kim, H., Jun, M.B. and Sutheralnd, J.W., 2019. Predictive Maintenance of Machine Tool Systems Using Artificial Intelligence Techniques Applied to Machine Condition Data. Procedia CIRP, 80, pp.506-511.

3] Cho, S., May, G., Tourkogiorgis, I., Perez, R., Lazaro, O., de la Maza, B. and Kiritsis, D., 2018, August. A Hybrid Machine Learning Approach for Predictive Maintenance in Smart Factories of the Future. In IFIP International Conference on Advances in Production Management Systems (pp. 311-317). Springer, Cham.

4] Liao, W., Pan, E. and Xi, L., 2010. Preventive maintenance scheduling for repairable system with deterioration. Journal of Intelligent Manufacturing, 21(6), pp.875-884.

5] Russell, S.J. and Norvig, P., 2016. Artificial intelligence: a modern approach. Malaysia; Pearson Education Limited,

6] Susto, G.A., Schirru, A., Pampuri, S., McLoone, S. and Beghi, A., 2014. Machine learning for predictive maintenance: A multiple classifier approach. IEEE Transactions on Industrial Informatics, 11(3), pp.812-820.

7] Li, H., Parikh, D., He, Q., Qian, B., Li, Z., Fang, D. and Hampapur, A., 2014. Improving rail network velocity: A machine learning approach to predictive maintenance. Transportation Research Part C: Emerging Technologies, 45, pp.17-26.

8] Allwood, C.M., 2012. The distinction between qualitative and quantitative research methods is problematic. Quality & Quantity, 46(5), pp.1417-1429.

9] Goldkuhl, G., 2012. Pragmatism vs interpretivism in qualitative information systems research. European journal of information systems, 21(2), pp.135-146.

10] Smith, J., Bekker, H. and Cheater, F., 2011. Theoretical versus pragmatic design in qualitative research. Nurse researcher, 18(2), pp.39-51.

11] Mihas, P., 2019. Qualitative data analysis. In Oxford Research Encyclopedia of Education.

12] McCusker, K. and Gunaydin, S., 2015. Research using qualitative, quantitative or mixed methods and choice based on the research. Perfusion, 30(7), pp.537-542.

13] Creswell, J.W. and Creswell, J.D., 2017. Research design: Qualitative, quantitative, and mixed methods approaches. Sage publications.

14] Connelly, L.M., 2014. Ethical considerations in research studies. Medsurg Nursing, 23(1), pp.54-56.

15] Furman, R., 2009. Ethical considerations of evidence-based practice. Social Work, 54(1), pp.82-84.

Domingos, P.M., 2012. A few useful things to know about machine learning. Commun. acm, 55(10), pp.78-87.

Jordan, M.I. and Mitchell, T.M., 2015. Machine learning: Trends, perspectives, and prospects. Science, 349(6245), pp.255-260.

Arel, I., Rose, D.C. and Karnowski, T.P., 2010. Deep machine learning-a new frontier in artificial intelligence research. IEEE computational intelligence magazine, 5(4), pp.13-18.

Russell, S.J. and Norvig, P., 2016. Artificial intelligence: a modern approach. Malaysia; Pearson Education Limited,.

Chen, M., Challita, U., Saad, W., Yin, C. and Debbah, M., 2017. Machine learning for wireless networks with artificial intelligence: A tutorial on neural networks. arXiv preprint arXiv:1710.02913.

Susto, G.A., Schirru, A., Pampuri, S., McLoone, S. and Beghi, A., 2014. Machine learning for predictive maintenance: A multiple classifier approach. IEEE Transactions on Industrial Informatics, 11(3), pp.812-820.

Li, H., Parikh, D., He, Q., Qian, B., Li, Z., Fang, D. and Hampapur, A., 2014. Improving rail network velocity: A machine learning approach to predictive maintenance. Transportation Research Part C: Emerging Technologies, 45, pp.17-26.

Kateris, D., Moshou, D., Pantazi, X.E., Gravalos, I., Sawalhi, N. and Loutridis, S., 2014. A machine learning approach for the condition monitoring of rotating machinery. Journal of Mechanical Science and Technology, 28(1), pp.61-71.

Obermeyer, Z. and Emanuel, E.J., 2016. Predicting the future—big data, machine learning, and clinical medicine. The New England journal of medicine, 375(13), p.1216.

Al-Jarrah, O.Y., Yoo, P.D., Muhaidat, S., Karagiannidis, G.K. and Taha, K., 2015. Efficient machine learning for big data: A review. Big Data Research, 2(3), pp.87-93.

Suthaharan, S., 2014. Big data classification: Problems and challenges in network intrusion prediction with machine learning. ACM SIGMETRICS Performance Evaluation Review, 41(4), pp.70-73.