Predictive Modeling of Failure States in Manufacturing Systems Using Artificial Intelligence in the Context of Sustainability

Machine Learning for Predictive Maintenance in Manufacturing

The findings reveal that ML algorithms significantly improve fault detection, Remaining Useful Life (RUL) estimation, and maintenance decision-making, but challenges related to data quality, model interpretability, cybersecurity, and integration with legacy systems continue to affect implementation effectiveness.

Optimising Aircraft Fleet Maintenance with Reinforcement Learning: A numerical case study

Unexpected failures and inefficient maintenance plans lead to costly downtime in many industrial fields, undermining operational sustainability and overall profitability, or, in the worst scenarios, creating unsafe situations for people. These issues justify a shift towards predictive maintenance, and structural health monitoring (SHM) has emerged as a viable option. By exploiting modern hardware and software technologies, the diagnosis and prognosis of monitored structures can be performed while accounting for estimation uncertainty. In addition, for complex structures, monitoring all components is not feasible for technical or economic reasons, and a mix of condition and schedule-based maintenance approaches is still required. To fully exploit the SHM benefits, an effective decision-making process is necessary, including Opportunistic Maintenance (OM) that accounts for both scheduled stops and the Residual Useful Life (RUL) predicted by prognostic models. In this context, decision-making becomes significantly more complex when managing an entire fleet rather than a single asset, reflecting the real-world challenges companies face. This study explores the application of Reinforcement Learning (RL) to automated decision-making in aircraft fleet maintenance, aiming to minimise operational and maintenance (O&M) life-cycle costs. The RL agent is trained and deployed within a fleet life cycle simulator that replicates real-world conditions, including mission schedules, maintenance operations, repair bases with varying capabilities, and potential unforeseen events. Each aircraft is modelled as an ensemble of subcomponents, with only a portion equipped with an SHM system that provides Remaining Useful Life (RUL) predictions, expressed as probability distributions, and fault detection. The agent leverages this data to develop an optimal maintenance policy, aligning scheduled stops for non-monitored components with necessary interventions for monitored parts. Specifically, it determines maintenance schedules and selects which components to replace based on fleet-wide data, expected mission schedules, and aircraft availability. While the study is demonstrated using a fleet life-cycle simulator, it serves as a preliminary step toward the real-world implementation of autonomous agents. The integration of the agent within the fleet simulator, potentially based on a real scenario, enhances its capabilities as a Digital Twin, supporting management decisions. Furthermore, it leverages the potential benefits of SHM technology, encouraging broader industry adoption.

A Data-Driven Method for Predicting The Operational Status of The Aircraft Assembly System

Aircraft assembly systems, as a critical phase in aerospace manufacturing, face significant challenges in maintaining production efficiency and ensuring product quality. This complex manufacturing system exhibits two distinct characteristics: (1) tightly coupled interactions among manufacturing elements involving process sequences, material flows, and equipment utilization; and (2) dynamic resource allocation and material distribution plans. The inherent variability in production element configurations often leads to operational instability and schedule deviations, which may result in abnormal production states. To address these challenges, this study proposes a data-driven predictive framework that integrates Long Short-Term Memory (LSTM) neural networks with multi-criteria evaluation. The developed LSTM-based model effectively forecasts two critical production indicators of cycle time and balance rate, achieving temporal prediction through historical operational data analysis. The proposed methodology facilitates timely anomaly detection and early warning, allowing proactive risk mitigation and ensuring sustained production system stability. This research contributes to advancing intelligent monitoring and control strategies for aircraft assembly operations within data-driven manufacturing environments.

Logical-probabilistic model for managing maintenance and repair processes of a shipbuilding company using copulas

Purpose of research . The purpose of research is to develop and substantiate a logic-probabilistic model for managing maintenance and repair processes of ship technical systems, taking into account the stochastic dependence between component failures using copula theory. The study also aims to quantitatively assess the impact of external operational factors on equipment reliability and on the accuracy of critical failure prediction Methods . The study applies the logic-probabilistic reliability analysis method (fault tree analysis) to formalize the structure of a ship system and describe the conditions leading to a critical system failure. To model statistical dependence between failures of interconnected components, the Clayton copula is employed, enabling the representation of strong lower-tail dependence. Failure probabilities of individual components are determined based on the exponential time-to-failure distribution. A corrective function reflecting the impact of exogenous factors (humidity, salinity and vibrational load) is incorporated into the model. Empirical validation is performed using operational data from 15 vessels of Project 22870 for the period 2020-2024, applying absolute and relative error metrics. Results . The results show the accounting for the dependence between the cooling pump and the generator increases the estimate annual probability of a critical system failure from 28.11% to 37.04% (a 31.77% relative increase). The probability of their joint annual failure reaches 17.52%, which is more than twice the estimate obtained under the assumption of statistical independence. Model validation demonstrates a deviation of less than 2.88% from observed data. Conclusion . The proposed model provides a more accurate assessment of operational failure risks and can serve as a tool for adaptive maintenance planning, as well as integration with digital twins and predictive maintenance systems in the shipbuilding industry.

Interchangeability and cannibalization in industrial maintenance: perspectives for asset management with machine learning, big data and artificial intelligence

Improving the Productivity of a Production Line through Explainable Artificial Intelligence for Data Analysis

Improving the performance of industrial production systems through data analysis is a recurrent topic in both research and industry. However, most existing data-driven approaches in manufacturing focus on engineer level decision support, such as predictive maintenance or strategic resource planning and pay little attention to operator-centred decision support. At the same time, production data exhibit specific characteristics that make their analysis challenging: high-dimensionality, often heterogeneous, and naturally structured as time series and point processes. These gathered properties require dedicated modelling strategies insufficiently addressed in the literature.An additional and increasingly recognised challenge is the adoption of digital decision-support tools on the shop floor. Operators and line adjusters need to trust the models’ outputs in order to integrate them into their daily practices. This calls for the design of explainable artificial intelligence approaches made for production contexts and for the needs of human operators.The proposed approach is developed in a real factory, to support operators decisions based on the state of the production system.. We detail the decision model building process from dimensional reduction to combination with field knowledge.Our goal is to build models that learn to recognise production situations similar to past ones that have already led to stops, so as to suggest relevant corrective actions to implement. A key aspect of our approach is the continuous confrontation of the modelling choices with field experience, in order to provide interpretable reasons for the recommendations.