Resilient Smart Manufacturing Systems Using Predictive Analytics and Digital Twin Technologies

Abstract

This study examined how predictive analytics and digital twin technologies contribute to resilient smart manufacturing systems in increasingly complex and disruption-prone industrial environments, where many firms still struggle with fragmented data systems, reactive maintenance, and limited adaptive response capability. The purpose of the research was to determine whether predictive analytics and digital twin technologies individually and jointly strengthen resilience outcomes in smart manufacturing settings. The study adopted a quantitative, cross-sectional, case-based design and drew on data from cloud-enabled and enterprise-oriented smart manufacturing cases involving production, operations, maintenance, quality, and digital systems personnel. A total of 240 questionnaires were distributed, 221 were returned, and 210 valid responses were used for final analysis, yielding an 87.5% valid response rate. The key independent variables were predictive analytics and digital twin technologies, while the dependent variable was resilient smart manufacturing systems, measured through dimensions such as disruption anticipation, adaptive response, recovery speed, operational continuity, and system flexibility. Data were collected using a five-point Likert scale instrument and analyzed through descriptive statistics, reliability testing, Pearson correlation, and multiple regression. The findings showed high mean scores for predictive analytics (M = 4.08, SD = 0.61), digital twin technologies (M = 3.96, SD = 0.66), and resilient smart manufacturing systems (M = 4.14, SD = 0.58). Correlation results revealed that predictive analytics had a strong positive relationship with resilience (r = 0.721, p < .001), while digital twin technologies also showed a strong positive relationship (r = 0.684, p < .001). Regression analysis indicated that predictive analytics (β = 0.482, p < .001) and digital twin technologies (β = 0.361, p < .001) significantly predicted resilience, with the overall model explaining 58.9% of the variance (R² = 0.589, F = 86.47, p < .001). The study implies that manufacturing firms can substantially improve operational continuity, adaptability, and recovery readiness by integrating predictive intelligence with digitally synchronized system representation, thereby advancing more resilient and strategically adaptive smart manufacturing ecosystems.