Smart Manufacturing Frameworks Using AI for Defect Prevention, Integrity Assurance, and Lifecycle Reliability in Energy Infrastructure

Abstract

AI-enabled smart manufacturing is increasingly adopted in energy-infrastructure production, yet organizations still face fragmented quality, inspection, maintenance, and reliability processes, creating uncertainty about which intelligent capabilities most effectively prevent defects, preserve asset integrity, and improve lifecycle performance. This study quantitatively evaluated the contribution of AI-enabled smart manufacturing capabilities to defect prevention, integrity assurance, and lifecycle reliability in energy-infrastructure manufacturing. A quantitative, cross-sectional, case-study-based design applied a five-point Likert-scale questionnaire to professionals in manufacturing, quality assurance, inspection, maintenance, reliability, automation, asset integrity, and industrial digital transformation. The study retained 312 valid responses from 360 distributed questionnaires, an 86.7% usable response rate. Key variables included AI-enabled predictive quality analytics, intelligent inspection and process monitoring, AI-enabled predictive maintenance, Industrial IoT and real-time data integration, digital-twin and AI-based process optimization, defect prevention, integrity assurance, and lifecycle reliability. Data were analyzed in SPSS using descriptive statistics, Cronbach’s alpha, Pearson correlation, and multiple regression. Reliability coefficients ranged from .821 to .914. Intelligent inspection and process monitoring recorded the highest technological mean (M = 4.02), while digital-twin optimization recorded the lowest (M = 3.71). Predictive quality analytics strongly predicted defect prevention (β = .361, p < .001), with the model explaining 54.8% of variance. Intelligent inspection was the strongest predictor of integrity assurance (β = .387, p < .001, R² = .596). Predictive maintenance was the strongest technological predictor of lifecycle reliability (β = .334, p < .001), and integrity assurance also contributed significantly (β = .291, p < .001); the lifecycle model explained 61.3% of variance. All eight proposed hypotheses received statistical support at the five percent level. Predictive quality, intelligent inspection, connected industrial data, digital twins, and predictive maintenance should be implemented as an integrated socio-technical system to strengthen manufacturing quality, asset integrity, and long-term energy-infrastructure reliability.