基于健康监测的南极望远镜在线故障检测与主动容错关键技术研究

The telescopes in Antarctic have high failure rate because of the extreme environment of Antarctic Dome A. Furthermore, the maintenance interval more than 1 years because of the unattended operation mode. Therefore, the telescope should have highly reliable. The main target of this project is to improve the reliability of telescope. The health prediction technology and active fault-tolerant control method are studied based on the telescope control system in the Antarctic. The main content of this project include: The innovative forecasting technology is put forward base on online monitoring data and historical data. The forecasting method use intelligent algorithm because of its nonlinear processing ability. The fault diagnosis base on health degradation of the telescope is proposed. The performance index is degradation threshold. The fault tolerance mechanism is proposed in order to improve the system reliability. The active fault tolerance method based on adaptive control law is put forward through the dynamics model of the control system. Finally, the control strategy is verified by load simulation and fault simulation. The results will lay a theoretical and technical basis for the long-term reliable operation of the Antarctic telescope or unattended large aperture telescope.

南极Dome A地区的极端环境导致望远镜故障率较高，而无人值守的运行特点使维修间隔达1年以上，因此，要求望远镜具有极高的可靠性。项目以提高望远镜可靠性为目标，以南极望远镜控制系统为背景，研究望远镜的健康预测技术和主动容错控制方法。包括：基于历史数据、在线监控数据，利用智能算法的非线性处理能力，创新性地提出用预测技术揭示极端环境的影响；以退化阈值为性能指标，融合南极望远镜的健康退化规律，提出基于健康退化的故障诊断模型；以提高系统可靠度为目标，通过控制系统的动力学模型，深入探索基于主动容错控制的容错机制，研究基于自适应控制律的主动容错方法；最后基于故障模拟及负载模拟的仿真及实验，验证控制策略的有效性。研究结果将为南极望远镜或无人值守大口径望远镜的长期可靠运行奠定理论和技术基础。

项目以提高南极望远镜可靠性为目标，围绕南极望远镜状态监测和故障诊断、南极台址参数预测和系统非线性补偿等问题开展研究。针对望远镜传动系统的状态监测，提出采用信号分解技术来提取信号的多域特征，结合特征融合和模态筛选技术，并利用智能算法的非线性处理能力来提升故障诊断的正确性。结合南极的历史台址监测数据，提出了利用机器学习方法，仅仅依靠南极Dome A的多层气象参数即可完成对视宁度的预测，为南极望远镜的观测策略的优化提供了新的工具和方法。以提高望远镜控制精度为目标，提出了采用模型预测控制的时滞补偿策略，对比传统的控制策略，控制精度有了大幅提升。