Quantifying the Impact of Cyber-attacks on IoT-Enabled Predictive Maintenance: Sensor Integrity, Model Performance, and Manufacturing Outcomes

Abstract

The increasing integration of the Internet of Things (IoT), industrial control systems, machine learning, and predictive maintenance has strengthened condition-based decision-making in smart manufacturing while simultaneously introducing cybersecurity risks capable of affecting both digital and physical operations. This quantitative study examined the effects of cybersecurity threats on sensor-data integrity, predictive-maintenance performance, equipment reliability, maintenance efficiency, and manufacturing outcomes in IoT-enabled predictive-maintenance environments. The final analytical dataset contained 186,420 observations, of which 143,740 (77.11%) represented normal operation and 42,680 (22.89%) represented cyber-compromised conditions. Cybersecurity scenarios included false-data injection, spoofing, replay, man-in-the-middle, denial-of-service, packet manipulation, communication disruption, and malware-related events. Results demonstrated that packet loss increased from 1.6% to 8.7%, communication latency increased from 24.6 ms to 71.8 ms, and mean signal deviation increased from 2.3% to 12.9% under cyber-compromised conditions. Mean predictive-classification accuracy declined by approximately 10.9 percentage points, while the false-negative rate increased from 6.8% to 18.9%. Remaining-useful-life prediction also deteriorated, with MAE increasing from 7.4 to 13.9 operating cycles and R² decreasing from 0.91 to 0.73. Operationally, cyber compromise increased failure frequency by 64.3% and unplanned downtime by 124.6%, while MTBF decreased by 38.6% and equipment availability declined from 96.2% to 89.1%. Throughput decreased by 18.5%, cycle time increased by 23.9%, and OEE declined from 86.9% to 74.1%. Multivariate analysis identified manipulated-data proportion, attack intensity, compromised-device exposure, equipment criticality, and legacy-system dependence as significant risk factors. Stronger cybersecurity controls reduced successful attacks from 43.6% to 18.7% and unplanned downtime by 49.4%. The findings established that cybersecurity integrity was quantitatively associated with predictive reliability, maintenance effectiveness, equipment availability, product quality, and overall smart-manufacturing performance.