Application of Predictive Maintenance Concepts Using Artificial Intelligence Tools

Abstract

The growing competitiveness of the market, coupled with the increase in automation driven with the advent of Industry 4.0, highlights the importance of maintenance within organizations. At the same time, the amount of data capable of being extracted from industrial systems has increased exponentially due to the proliferation of sensors, transmission devices and data storage via Internet of Things. These data, when processed and analyzed, can provide valuable information and knowledge about the equipment, allowing a move towards predictive maintenance. Maintenance is fundamental to a company’s competitiveness, since actions taken at this level have a direct impact on aspects such as cost and quality of products. Hence, equipment failures need to be identified and resolved. Artificial Intelligence tools, in particular Machine Learning, exhibit enormous potential in the analysis of large amounts of data, now readily available, thus aiming to improve the availability of systems, reducing maintenance costs, and increasing operational performance and support in decision making. In this dissertation, Artificial Intelligence tools, more specifically Machine Learning, are applied to a set of data made available online and the specifics of this implementation are analyzed as well as the definition of methodologies, in order to provide information and tools to the maintenance area.

Acoustic Anomaly Detection of Mechanical Failures in Noisy Real-Life Factory Environments

“…The results of this study suggest what acoustic detection of failures could be used for Predictive Maintenance [61] of industrial machinery in the context of Industry 4.0. The incorporation of acoustic new sensor technologies combined with deep learning methods can be used to avoid premature replacement of equipment, saving maintenance costs, improving machining process safety, increasing availability of equipment, and maintaining the acceptable levels of performance [2].…”

Study and Application of Industrial Thermal Comfort Parameters by Using Bayesian Inference Techniques

“…In this sense, Diogo Cardoso and Luís Ferreira mention that "artificial intelligence tools, specifically automated learning, show great potential for analyzing great amounts of data, now readily available, to reduce upkeeping costs and increase operational performance and decision-making support" [50].…”

Systems Engineering: Availability and Reliability

“…Artificial Intelligence tools, in particular Machine Learning, exhibit enormous potential in the analysis of large amounts of data, now readily available, thus aiming to improve the availability of systems, reduce maintenance costs and increase operational performance and support in decision making. This is why Cardoso and Ferreira [7] apply Machine Learning to a set of data made available online and the specifics of this implementation are analyzed, as well as the definition of methodologies, in order to provide information and tools to the maintenance area. Although the results obtained compare well with those presented so far in the literature, the biggest disadvantage in using the presented methodology lies in the definition of the features.…”

“…If the selection of features is not the most correct, the results obtained can lead to wrong predictions. For future work, the application of feature learning concepts will be considered instead of feature engineering, which appears to be promising to improve the results obtained [7]. This section may be divided into subheadings.…”