DSpace-CRIS

Due to the global uncertainty caused by social problems such as COVID-19, the microchip crisis, and cybersecurity attacks, many companies have opted to use emerging technologies since they allow them to maintain productivity in challenging times, achieving business strategic results. The preceding is based on market studies and research on emerging technologies. One example is in wearable artificial intelligence (AI) and wireless sensor networks (WSNs), the application of which in Industry is growing exponentially. Some social trends that seek to explore the use of AI and sensors to solve these social problems and uncertainty are Society 5.0 (S5.0), Industry 5.0 (I5.0), as well as the so-called AI for Social Good (AI4SG). I5.0 seeks to improve the resilience of industrial processes through sustainable and human-centered strategies; resilience is a metric closely related to the maintenance of physical assets since it can be calculated through performance and time measures in calculating reliability, availability, and safety. To have an approach to the problems of industrial maintenance and to be able to formulate a research question, first maintenance congresses, forums, and social networks were attended, an own initiative of Industry 4.0 was created, and experts were interviewed. The above actions allowed us to understand the industrial maintenance needs and problems more clearly. Secondly, an explorative literature review (ELR) was carried out on the research gaps in industrial maintenance; the gaps that were most consistent with the interest of this research were the following: a) the creation of feedback mechanisms for the state of physical assets; b) intelligent human-machine interaction, synergy, interoperability, and mutual learning; and c) resolve the issue of uncertainty. Third, a comparative table was then made in which the Industry's needs were compared against the selected research gaps. Finally, this comparison helped to justify the research topic: A human-in-the-Loop-based Framework for Industrial Physical Assets Resilience. Intelligent maintenance systems, such as reliability-centered maintenance (RCM), maintenance 4.0 (M4.0), self-healing systems, condition-based maintenance (CBM), and risk-based maintenance, among others, can repair a machine or take care of a physical asset; however, if the automatic and self-repairing system fails or decides out of context, the physical asset is left partially or unprotected, increasing uncertainty in the productive system and requiring external help. Continuous collaboration and communication between machines and workers are necessary to keep physical assets running in an industrial process. Due to the above, one motivation of this research work is to know how including human decisions in intelligent maintenance systems can impact the increase of resilience in physical assets. By including the human-in-the-loop (HITL), the maintenance worker could continue to put his knowledge at the service of the physical assets, which means that he keeps his practical ability to solve problems. Because the above is not considered in other maintenance frameworks, such as Maintenance 4.0, the following main research question (RQ1) has been formulated: could a human-in-the-loop-based maintenance framework improve the resilience of physical assets? This doctoral research examines how industrial physical assets can be protected to increase their resilience by creating a novel Maintenance 5.0 (M5.0) framework with the HITL. Due to the above, a framework for the worker of the future called Worker 5.0 (W5.0) is presented, which is the human being who adds value to the company's value chain through the use of non-intrusive wearables that increase their communication, and through the experience, it shares advising on the AI, so that the decisions that are made do not get out of context; the preceding impacts on an increase in the resilience of the signal that the physical asset is transmitting. Another literature review was conducted, but this time it was a systematic literature review (SLR), taking as keywords: HITL, maintenance, and resilience; with the SLR results, the state of the art was elaborated, and the hypothesis was formulated. With the SLR results and applying a questionnaire survey to experts, Checkland's qualitative and deductive methodology was selected to design a maintenance framework that meets the objectives of Industry 5.0. Within this methodology, to know the goals and answer of the research question, some specific activities were carried out: a) detection of needs and review of the scope of research gaps in maintenance; b) a characterization of work and maintenance throughout industrial history; c) definitions and characteristics of M5.0 and W5.0 are proposed, d) two HITL control loops of M5.0 framework are proposed: the first is an OSA-CBM-inspired, and the second is self-healing-inspired; e) M5.0 general architecture is proposed, which considers the OODA loop, as well as the steps for the creation of value in the company and classifies the activities that must be carried out in the physical world and the cyber world; f) two proposals for the calculation of resilient maintenance: one for real- time processes, inspired by the raised cosine, and another for historical processes, inspired by the Weibull distribution; g) an ease-benefit analysis of the proposed changes to M4.0 to make it M5.0; h) a list of enabling technologies for the proposed maintenance framework noting their impact on I5.0 goals; i) a catalog of wearables was prepared that will help implement W5.0, and M5.0; and finally, j) tools and enablers to implement M5.0 are proposed. In the last step of the methodology, the theoretical proposal must be carried out in the physical world, so the hypothesis is demonstrated through a case study of an Industrial Wireless Sensor Network (IWSN), which refers to a Wireless Sensor Network (WSN) that is immersed in an industrial environment. One of the leading research gaps was the creation of feedback mechanisms of the state of physical assets; at the same time, one of the needs detected in the Industry was the use of wireless and non-intrusive technologies (wearables and IWSNs) to implement remote solutions; due to this, to validate the proposed M5.0 framework, it was decided to carry out a case study, taking an IWSN as a physical asset. The maintenance of the IWSN may vary according to first, a) the inclusion of HITL, which is the primary motivation of this research, and responds to RQ1, and second, b) the improvement in communication between network nodes (physical assets) that make up the IWSN, which responds to RQ2: which Industrial Wireless Sensor Network scheme helps more to protect the network nodes from failure events? IWSNs are physical assets that must be maintained at acceptable levels since they are the ones that make up the base of the automation pyramid, feeding data to industrial control systems. RQ2 is considered a sub-item of RQ1 since it responds to the needs of this case study where M5.0 is being validated. To answer RQ2, a mechanism for the resilience of IWSNs that is based on the measurement of performance metrics and the network scheme is proposed, laying the foundations of the case study; the case study is composed of simulations and experiments that use different densities and network schemes; while the cooperative scheme is more efficient when an IWSN is under normal operating conditions, the collaborative scheme offers more excellent protection against aggressive interference on performance metrics, making it more secure and resilient. In addition, a real-time jamming detection algorithm is proposed with the following characteristics: first, a) it examines the characteristics and damages caused by the type of aggressor; second, b) it reflects the natural immunity of the WSN (which depends on its node density and a cooperative or collaborative configuration); and finally, c) it considers performance metrics, especially those that impact power consumption during sensor transmission. The case study demonstrated that network nodes save more than 10% of energy if they switch their network schemes in the event of interference. Regarding jamming mitigation, it is possible to know the optimal route for delivering communication packets using an AI algorithm; M5.0 is finally validated by including the HITL as a network adviser, which assesses AI algorithms to increase resilience.

Tesis

Browse

Filters

Settings

Sort By

Results per page

Search Results