摘要：当前涡扇发动机加减速控制计划,不能规划双轴加速进程,导致加速末期推力波动。本文提出了分阶段的加速度控制计划,按边界约束划分加速进程,基于动态稳定法设计初始加速度计划,通过寻找系统参数的超调临界点,协同修正初始加速度计划,满足边界约束的同时保证双转子加速时间相同。在仿真验证环节,与基于定状态法的控制计划比较,本文方法可避免燃油振荡现象,推力超调量降低2.43%,推力稳定时间缩短1.045 s;与等转速增量的加速度控制计划比较,进一步改进了本文方法,推力响应时间缩短0.24 s,控制计划设计耗时为7.5781 s。综上,本文方法缩短了推力超调量和推力稳定时间,有望应用于发动机控制计划在线计算。

摘要：Saturator is one of the core components of humid air turbine （HAT） and is the main feature of HAT making it different from other gas turbine cycles. Due to the lack of sufficient experience in commercial plant operation, HAT cycle has a great demand for modeling and simulation of the system and its components, especially the saturator, to provide reference for system design and optimization. The conventional saturator models are usually based on the theory of heat and mass transfer, which need two accurate coefficients to ensure convincing results. This work proposes a global heat and mass transfer coefficient based on cooling tower technology to model the saturator in small-scale HAT cycle. Compared with the experimental data, the simulation results show that the proposed model well predicts the dynamic humidity and temperature distribution characteristics of saturator at low air pressure and temperature.

摘要：得益于机器学习理论及技术的高速发展,数据驱动的预测及健康管理(Prognostic and Health Management, PHM)系统能从高维的多源异构数据中学习设备不同状态的特征和趋势,进而完成诊断和预测任务。然而,复杂机器学习模型的不透明性也引发了使用者对模型的信任问题,且设计者难以根据自身知识针对性改进模型。为此,本文总结了机器学习可解释性研究进展,归纳了其研究范式,概述了模型可解释性在PHM领域中的应用,并对比了当前可解释性研究在机器学习和PHM中的研究模式,提出未来的一些研究设想。

摘要：提出了机理模型和基于神经网络的数据驱动模型相结合的燃机混合模型结构,利用燃机实际运行数据,通过数据驱动方法修正由于特性线偏差导致的机理模型仿真误差。以天然气长输管道压气站的某型双轴燃气轮机为研究对象,设计了3种机理-数据混合模型结构。通过对比3种混合模型与机理模型的性能仿真结果,说明与燃机机理模型相比,混合模型精度相对更高。同时基于燃机整体并联补偿的混合模型结构对截面参数的仿真精度最高,各参数的平均相对误差基本在1. 5%以内,其中压气机出口和动力透平出口温度的相对误差小于0. 5%。由于数据驱动方法可以针对特定燃机利用运行数据训练出特定的补偿网络,对于不同的机组所设计的混合结构模型都能具有较高的准确性,为提高燃机模型精度提供了参考手段。

摘要：Due to the fact that the turbine outlet temperature of aeroderivative three-shaft gas turbine is low,the conventional combined cycle is not suitable for three-shaft gas ***,the humid air turbine(HAT)cycle provides a new choice for aeroderivative gas turbine because the humidification process does not require high *** HAT cycle plants are all based on single-shaft gas turbines due to their simple structures,therefore converting aeroderivative three-shaft gas turbine into HAT cycle still lacks sufficient *** paper proposes a HAT cycle model on a basis of an aeroderivative three-shaft gas *** HAT cycle modelling of saturator,gas turbine and heat exchanger are carried out based on the modular modeling *** models are verified by simulations on the aeroderivative three-shaft gas *** results show that the studied gas turbine with original size and characteristics could not reach the original turbine inlet temperature because of the introduction of ***,the efficiency still increases by 0.16%when the HAT cycle runs at the designed power of the simple ***,simulations considering turbine modifications show that the efficiency could be significantly *** results obtained in the paper can provide reference for design and analysis of HAT cycle based on multi-shaft gas turbine especially the aeroderivative gas turbine.

摘要：燃气轮机的传感器经常工作在高温、振动、腐蚀等恶劣环境下,传感器的故障难以避免且影响巨大。控制图是一种实用的生产过程产品质量管理工具。然而,燃气轮机的传感器监测数值受机组工况,大气环境、机组健康状况众多因素影响,难以直接通过控制图进行故障诊断。本文基于性能仿真技术,对传统控制图进行改造,提出一种变控制线的控制图,并设计了一套使用该控制图进行燃气轮机传感器故障诊断的流程,可以并行评价机械设备和传感器的健康状况。最后对一台航改型工业燃气轮机应用该方法进行传感器故障诊断案例研究,结果表明该方法可以有效地识别出故障传感器,并进行数据恢复,减少非计划停机,改善燃气轮机运行状况。

摘要：利用燃气轮机性能仿真模型,分析燃气轮机效率、预热器成本和污染物排放量对燃气轮机经济性的影响。针对某型号单轴燃气轮机,设计一个燃料预热器。考虑压损,仿真出效率随燃料温度的变化,拟合出效率、预热器成本和污染物排放量与预热器等效换热面积的关系,并将其转化为燃气轮机运行总收支,分析燃气轮机的经济性。结果表明,当等效换热面积为1 m2,燃料温度为180℃时,案例中燃气轮机节省的收益最多,为15万元。在保证燃气轮机安全、可靠性的前提下,优化运行收支,可实现燃气轮机经济效益的最大化。