This research contributes to the field of reliability engineering and system safety by introducing aninnovative diagnostic method to enhance the reliability and safety of complex technological systems. Steam turbines are specifically referred to. This study focuses on the integration of advanced signal processing techniques and engineering dynamics in addressing critical issues in the monitoring and maintenance of mechanical systems. By utilizing genetic algorithms, we improve the capability to detect, localize, and ascertain the causes of both singular and intricate degradations, including three-fold and four-fold faults, within steam turbine operations. We can detect degradation with accuracies of 72.6% for three-fold faults and 62.2% for four-fold faults. This significant advancement emphasizes the potential for improved machine and structural health monitoring, especially where non-stationary and random vibrations are common, such as in powertrain and drivetrain systems. This methodology is vital for the maintenance and operational strategies of critical infrastructures like nuclear power plants, chemical plants, and manufacturing facilities where steam turbines play a crucial role. The novelty of this approach lies in the use of genetic algorithms for thermal-flow diagnostics of steam turbines, which had been unaddressed in literature. Moreover, the merger of theoretical and experimental aspects in this study underscores its relevance to practical applications, thereby demonstrating an original contribution to engineering knowledge and showcasing significant advancements over established methods. The research underscores the method’s potential as a universal tool for diagnosing complex systems, representing an advance in reliability engineering practices. By applying genetic algorithms, a noticeable link to improving the safety and reliability of technological systems is established, offering valuable insights into the design, maintenance, and extension of the lifespan of critical infrastructure.

中文翻译：

---

基于遗传算法的高级信号处理，用于蒸汽轮机中的多重退化检测


本研究通过引入一种创新的诊断方法来提高复杂技术系统的可靠性和安全性，为可靠性工程和系统安全领域做出了贡献。特别提到了蒸汽轮机。本研究的重点是将先进的信号处理技术和工程动力学相结合，以解决机械系统监测和维护中的关键问题。通过利用遗传算法，我们提高了在蒸汽轮机运行中检测、定位和确定单一和复杂退化原因的能力，包括三重和四重故障。我们可以检测到退化，三重故障的准确率为 72.6%，四重故障的准确率为 62.2%。这一重大进步强调了改进机器和结构健康监测的潜力，尤其是在非平稳和随机振动很常见的情况下，例如在动力总成和传动系统中。这种方法对于蒸汽轮机起着关键作用的关键基础设施（如核电站、化工厂和制造设施）的维护和运营策略至关重要。这种方法的新颖之处在于使用遗传算法对蒸汽轮机进行热流诊断，这在文献中一直没有涉及。此外，本研究中理论和实验方面的融合强调了其与实际应用的相关性，从而展示了对工程知识的原始贡献，并展示了对现有方法的重大进步。该研究强调了该方法作为诊断复杂系统的通用工具的潜力，代表了可靠性工程实践的进步。 通过应用遗传算法，建立了与提高技术系统安全性和可靠性的显着联系，为关键基础设施的设计、维护和延长使用寿命提供了有价值的见解。