WO2019144384A1 - 一种航空发动机启动过程排气温度预测方法 - Google Patents

一种航空发动机启动过程排气温度预测方法 Download PDF

Info

Publication number
WO2019144384A1
WO2019144384A1 PCT/CN2018/074353 CN2018074353W WO2019144384A1 WO 2019144384 A1 WO2019144384 A1 WO 2019144384A1 CN 2018074353 W CN2018074353 W CN 2018074353W WO 2019144384 A1 WO2019144384 A1 WO 2019144384A1
Authority
WO
WIPO (PCT)
Prior art keywords
prediction
data
starting process
exhaust temperature
aircraft engine
Prior art date
Application number
PCT/CN2018/074353
Other languages
English (en)
French (fr)
Inventor
汪锐
刘敏
张硕
李济邦
Original Assignee
大连理工大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 大连理工大学 filed Critical 大连理工大学
Priority to PCT/CN2018/074353 priority Critical patent/WO2019144384A1/zh
Priority to US16/629,893 priority patent/US20200234165A1/en
Publication of WO2019144384A1 publication Critical patent/WO2019144384A1/zh

Links

Classifications

    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02CGAS-TURBINE PLANTS; AIR INTAKES FOR JET-PROPULSION PLANTS; CONTROLLING FUEL SUPPLY IN AIR-BREATHING JET-PROPULSION PLANTS
    • F02C9/00Controlling gas-turbine plants; Controlling fuel supply in air- breathing jet-propulsion plants
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00Computing arrangements based on specific mathematical models
    • G06N7/01Probabilistic graphical models, e.g. probabilistic networks
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F02COMBUSTION ENGINES; HOT-GAS OR COMBUSTION-PRODUCT ENGINE PLANTS
    • F02CGAS-TURBINE PLANTS; AIR INTAKES FOR JET-PROPULSION PLANTS; CONTROLLING FUEL SUPPLY IN AIR-BREATHING JET-PROPULSION PLANTS
    • F02C7/00Features, components parts, details or accessories, not provided for in, or of interest apart form groups F02C1/00 - F02C6/00; Air intakes for jet-propulsion plants
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00Computer-aided design [CAD]
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00Machine learning
    • G06N20/20Ensemble learning
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05DINDEXING SCHEME FOR ASPECTS RELATING TO NON-POSITIVE-DISPLACEMENT MACHINES OR ENGINES, GAS-TURBINES OR JET-PROPULSION PLANTS
    • F05D2260/00Function
    • F05D2260/81Modelling or simulation
    • FMECHANICAL ENGINEERING; LIGHTING; HEATING; WEAPONS; BLASTING
    • F05INDEXING SCHEMES RELATING TO ENGINES OR PUMPS IN VARIOUS SUBCLASSES OF CLASSES F01-F04
    • F05DINDEXING SCHEME FOR ASPECTS RELATING TO NON-POSITIVE-DISPLACEMENT MACHINES OR ENGINES, GAS-TURBINES OR JET-PROPULSION PLANTS
    • F05D2270/00Control
    • F05D2270/01Purpose of the control system
    • F05D2270/11Purpose of the control system to prolong engine life
    • F05D2270/112Purpose of the control system to prolong engine life by limiting temperatures

Definitions

  • the invention belongs to the technical field of aircraft engine prediction, and particularly relates to a method for predicting exhaust gas temperature during aero engine starting process.
  • model-based methods When the aircraft is started, the aero-engine is in a state of high temperature, high load and high speed. The possibility of over-temperature of the engine is high, which increases the risk of flight. Therefore, it is necessary to predict the exhaust gas temperature and control the aeroengine. Overheated.
  • model-based methods There are three main methods for predicting aeroengine exhaust temperature: model-based methods, regression-based methods, and machine learning-based methods.
  • the model-based method is computationally complex, and the real-time calculation may have problems such as iterative convergence.
  • the regression-based method sometimes does not necessarily have obvious linear or other functional relationships between variables, and the model is difficult to select; the machine learning-based method has very Strong nonlinear mapping ability, short training time.
  • the technical problem to be solved by the present invention is to fill the gap of the exhaust gas temperature prediction during the starting process of the aeroengine, and to provide a method for predicting the exhaust gas temperature during the aeroengine starting process, using the aircraft engine ground. ⁇ 0 2019/144384 ⁇ (:17 € ⁇ 18/074353
  • the interview vehicle data is obtained by machine learning method to obtain the exhaust gas temperature prediction model of the engine starting process.
  • the model has high prediction accuracy and good generalization performance, and the prediction result can be further used for engine control.
  • an exhaust gas temperature prediction method for an aeroengine starting process is provided, and the technical solution of the present invention is as follows:
  • the data of the aircraft engine interview vehicle collected by the sensor such as high-pressure compressor speed, low-pressure compressor speed, main fuel flow, low-pressure turbine temperature, etc.
  • a parameter with a large correlation with the exhaust gas temperature is selected as an input parameter by a suitable correlation method.
  • phase space reconstruction is performed on the selected parameters to construct input and output samples.
  • the machine learning algorithm is used to predict the exhaust gas temperature, and the exhaust gas temperature prediction model of the aeroengine starting process is obtained with high prediction accuracy, high generalization ability and good robustness.
  • the identification of the abnormal data is performed using a density-based method and then culled.
  • the smoothing of the data uses a special function smoothing method.
  • the correlation analysis method uses the mutual information method.
  • the integrated algorithm integrates the Weak Learning Machine Extreme Learning Machine ( ELM ) to obtain a powerful learning machine with superior effects.
  • ELM Weak Learning Machine Extreme Learning Machine
  • the invention has the advantages of high prediction precision, strong generalization ability and strong robustness, and can predict the exhaust temperature of the aeroengine starting process in real time, and the prediction result can be used for the control of the aeroengine, etc.
  • the present invention uses the fusion prediction to include more information, so that the prediction error is reduced.
  • the invention integrates the weak learning machine by using the AdaB00St.RT integration algorithm. The prediction error is smaller.
  • Figure 1 is a flow chart of the present invention.
  • the data of the interview vehicle data of the aero engine starting process collected by the sensor is
  • ⁇ 3 ⁇ 4 ⁇ is the corresponding time series
  • N is the number of parameters
  • n is the number of samples.
  • the pre-processing data of the interview vehicle during the aero engine starting process includes the identification and processing of abnormal points, data smoothing processing and data normalization processing.
  • the Euclidean distance can be expressed as:
  • the point 3 ⁇ 4 is an abnormal point of the time series ⁇ , where is the upper limit of the abnormal factor.
  • the identified abnormal points are culled and filled with the mean of the adjacent data values at the left and right ends.
  • the second-order exponential smoothing method is used to process the abnormal point identification and processed data / ⁇ to remove noise or data pollution that may occur during signal acquisition.
  • / ⁇ 1 is the corpse ⁇ abnormal point identification and processed aeroengine performance parameter data
  • ⁇ ⁇ ⁇ is x ⁇ abnormal point identification and processed time series.
  • the second exponential smoothing algorithm is as follows:
  • 3 ⁇ 4 (1) 3 ⁇ 4 1 +(1- «)3 ⁇ 4-1 (1) (8)
  • the smoothed data / ⁇ 2 is normalized and converted into data in the range [ 0,1 ], where
  • exhaust gas temperature £010 is used as an input parameter for the prediction model.
  • phase space reconstruction is performed on the one-dimensional time series data in order to fully reveal the hidden information.
  • the number is the delay time, which is solved by the mutual information method and the ⁇ 10 method respectively.
  • the input and output samples are constructed according to the phase space reconstruction, as shown in Table 1 , where / is the prediction step size.
  • the present invention uses the AdaBoost.RT_ELM algorithm to predict the exhaust temperature of an aeroengine starting process.
  • the specific algorithm of AdaBoost.RT_ELM is as follows: ⁇ 0 2019/144384 ⁇ (:17 € ⁇ 18/074353
  • Step 4 Update the sample weight Z) f :
  • is the normalization factor.
  • the invention adopts an extreme learning machine ( ELM ) with fast learning speed and good generalization performance as a weak learning machine, and obtains a strong learning machine with high prediction precision by setting an appropriate number of iterations and a threshold value, that is, an exhaust temperature of an aeroengine starting process. Forecast model.
  • ELM extreme learning machine

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • General Physics & Mathematics (AREA)
  • Combustion & Propulsion (AREA)
  • Chemical & Material Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Mechanical Engineering (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Mathematical Analysis (AREA)
  • Computational Mathematics (AREA)
  • Pure & Applied Mathematics (AREA)
  • Algebra (AREA)
  • Probability & Statistics with Applications (AREA)
  • Mathematical Optimization (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Combined Controls Of Internal Combustion Engines (AREA)

Abstract

本发明属于航空发动机预测技术领域,提供了一种航空发动机启动过程排气温度预测方法。本发明所要解决的技术问题是填补航空发动机在启动过程中排气温度预测的空缺,提供一种航空发动机启动过程排气温度预测方法,利用航空发动机地面试车数据,采用机器学习的方法得到发动机启动过程排气温度预测模型,该模型预测精度高、泛化性能好,预测结果可进一步用于发动机控制方面等,降低了发动机出现超温的可能性。相对于传统的单参数预测,本发明由于采用融合预测,包含了更多信息,使得预测误差降低;相对于单一的预测算法,本发明由于采用了AdaBoost.RT集成算法,对弱学习机进行集成,预测误差更小。

Description

\¥0 2019/144384 卩(:17€謂18/074353
1
一种航空发动机启动过程排气温度预测方法
技术领域
本发明属于航空发动机预测技术领域, 具体涉及一种航空发动机启动过程 排气温度预测方法。
背景技术
在飞机启动时, 航空发动机处于高温、 高负荷和高转速的状态, 发动机出 现超温的可能性较大, 增加了飞行风险, 因此, 需要对排气温度进行预测以及 时对航空发动机进行控制防止出现超温。 对航空发动机排气温度预测的方法主 要有三种: 基于模型的方法、 基于回归的方法和基于机器学习的方法。 其中, 基于模型的方法计算复杂、 实时计算可能出现迭代不收敛等问题; 基于回归的 方法有时变量之间不一定有明显的线性或者其他函数关系, 模型很难选择; 基 于机器学习的方法具有非常强的非线性映射能力, 训练时间短。 在文献《基于 支持过程向量机的航空发动机排气温度预测》 中, 于广斌等人提出一种支持过 程向量机模型, 并运用到航空发动机排气温度预测中以预测航空发动机气路性 能衰退规律, 预测精度高。在文献《Application of Neural Networks in Forecasting Engine Systems Reliability》 中, Xu K等人利用神经网络对航空发动机排气温度 进行预测以预测发动机系统故障和可靠性。 以上方法都是以航空发动机多个飞 行循环的排气温度为数据进行预测, 以显示航空发动机的性能状态及退化情况, 但没有涉及到整个启动过程的排气温度预测, 因此无法在有超温现象前对发动 机进行控制。
发明内容
本发明所要解决的技术问题是填补航空发动机在启动过程中排气温度预测 的空缺, 提供一种航空发动机启动过程排气温度预测方法, 利用航空发动机地 \¥0 2019/144384 卩(:17€謂18/074353
2
面试车数据, 采用机器学习的方法得到发动机启动过程排气温度预测模型, 该 模型预测精度高、 泛化性能好, 预测结果可进一步用于发动机控制方面等。
按照本发明, 提供了一种航空发动机启动过程排气温度预测方法, 本发明 的技术方案如下:
首先, 对传感器采集到的航空发动机地面试车数据, 如高压压气机转速、 低压压气机转速、 主燃油流量、 低压涡轮后温度等进行预处理, 主要包括对异 常数据的识别和处理、 数据的平滑处理及数据的归一化处理。 然后, 基于信息 融合的思想, 通过合适的相关性方法选择与排气温度相关性大的参数作为输入 参数对其进行预测。 另外, 对选择的参数进行相空间重构构建输入输出样本。 最后, 采用机器学习的算法对排气温度进行预测, 获得预测精度高、 泛化能力 强、 鲁棒性好的航空发动机启动过程排气温度预测模型。
优选地, 在本发明中, 对异常数据的识别采用基于密度的方法然后进行剔 除。 对数据的平滑处理采用专用函数平滑法。 相关性分析方法采用互信息法。
Figure imgf000004_0001
集成算法, 通过对弱学习机极限学习机 (ELM) 进行集成, 得到效果优越的强 学习机。
本发明的有益效果: 本发明的预测模型预测精度高、 泛化能力强、 鲁棒性 强, 可对航空发动机启动过程的排气温度进行实时预测, 预测结果可用于航空 发动机控制等方面, 降低了发动机出现超温的可能性。 相对于传统的单参数预 测, 本发明由于采用融合预测, 包含了更多信息, 使得预测误差降低; 相对于 单一的预测算法, 本发明由于采用了 AdaB00St.RT集成算法, 对弱学习机进行 集成, 预测误差更小。
附图说明 \¥0 2019/144384 卩(:17(:\2018/074353 图 1是本发明的流程图。
具体实施方式
为了使本发明的目的、 技术方案及优点更加清楚明白, 以下结合附图及技 术方案, 对本发明进行进一步详细说明。
一、 航空发动机启动过程地面试车数据预处理
设传感器采集到的航空发动机启动过程地面试车数据 Data
Data = [Parax , Para2,…, Para!,…, ParaN ] ( 1 )
Parat ={xu}^=1 ,1 = 1,2,· - -,N ( 2)
Figure imgf000005_0001
等航空发动机性能参数数据, {¾} 为相对应的时间序列, N为参数个数, n为 样本个数。航空发动机启动过程地面试车数据预处理包括异常点的识别与处理、 数据平滑处理及数据归一化处理。
1、 基于密度的异常点识别
在时间序列{¾};1中, 点¾临近点的数量越少, 即其周围点密度较小, 则 点¾就更加可能为一异常点。 对于时间序列丨^} 中任意两点形成的有效点对 (¾,¾), 其欧几里得距离可表示为:
Figure imgf000005_0002
对有效点对 (¾,¾)中的点 ¾,定义其/:近邻点距离 ( >0, £^)为/:- (¾ ), 且 - (¾ )满足:
Figure imgf000005_0003
2)在{¾}^=1中, 满足 (¾, ) < (¾)的数据点的数量至多为 1个。 对有效点对(¾.,¾.)中的点¾, 称
厂一^4 (XII , XII) = rmx(dist(xli,:^·), 一 (:¾)) (4) \¥02019/144384 卩(:17€謂18/074353
4
为有效点; ¾¾的 近邻点极限距离。
为衡量¾点周围点数量的多少, 定义局部极限密度这一概念, 称
Figure imgf000006_0001
为点 ¾的 /:局部极限密度, 其中 (¾.)为点 的/:近邻点集合。
Figure imgf000006_0002
³ 则点¾为时间序列^ 的一个异常点, 其中 为异常因子上限。
对识别出的异常点进行剔除, 并用左右两端相邻数据值的均值填充。
2、 基于二次指数平滑法的数据平滑处理
采用二次指数平滑法对异常点识别及处理后的数据/^ 进行处理, 以去除 在信号采集过程中可能出现的噪声或数据污染, 其中
Figure imgf000006_0003
/^ 1为尸 ^异常点识别及处理后的航空发动机性能参数数据, {^ ^为 x } 异常点识别及处理后的时间序列。
二次指数平滑算法如下:
¾(1)1+(1-«)¾-1(1)
Sli {2)=aSli m +(l-a)Sli-l {2) 其中, 《为平滑系数; ¾ ¾ 分别为一次和二次平滑值, 定义平滑初始值 〇 为
^〇 = _ å ·1 (9) /=1
3、 数据归一化处理 \¥0 2019/144384 卩(:17€謂18/074353
5
对平滑处理后的数据/^2进行归一化处理, 转化为[0,1]范围内的数据, 其 中
Figure imgf000007_0001
计算各个参数与排气温度的互信息函数值, 考虑到各个参数与排气温度相 关性大小的差异及训练预测模型的时间要求, 取互信息函数值最大的三个参数 \¥0 2019/144384 卩(:17€謂18/074353
6
及排气温度 £010作为预测模型的输入参数。
三、 航空发动机启动过程地面试车数据相空间重构
因为航空发动机启动过程地面试车数据是一组时间序列数据, 为把其隐含 的信息充分显露出来, 对一维时间序列数据进行相空间重构。 对相关性分析后 的数据 进行相空间重构, 其中
Figure imgf000008_0001
/¾厂¾3 = |¾3 \=1,1 = 1,2, 3, 4 ( 19 ) 特别地, 尸麵 4 3 = ^3。 对于时间序列^^3:^, 对其重构后的相空间为
= [¾,¾,…, ¾,…,¾ ] ( 20) 其中,
¾ = [¾, Xi(I+t), - -, ¾(/+(«-!>) ], / = 1,2,…, ; = « - ( - 1>· ( 21 ) 为嵌入维数, 是延迟时间, 分别用互信息法和〇10法进行求解。根据相空间 重构构造输入输出样本, 如表 1所示, 其中 /为预测步长。
1 基于相空间重构的输入输出数据
Figure imgf000008_0002
四、 航空发动机启动过程排气温度预测模型
本发明采用 AdaBoost.RT_ELM算法对航空发动机启动过程的排气温度进行 预测。 AdaBoost.RT_ELM具体算法如下: \¥0 2019/144384 卩(:17€謂18/074353
7
(1)输入
Figure imgf000009_0001
样本和预测错误的样本。
2)初始化
令初始迭代次数?=1 ;
令第一次训练时, 训练样本权值分布为^(/)=1^,/=1,···^;
令初始误差率 =0
(3)迭代过程
/〇厂 1 = 1, , 1 :
Figure imgf000009_0002
Step 3:设置 /?f =<, a可以为 1,2或 3;
Step 4:更新样本权值 Z)f :
Figure imgf000009_0003
其中^是标准化因子。 通过调整各样本的权值, 即增加预测误差大的样本的权 值, 减少预测误差小的样本的权值, 使得误差大的样本在下一次迭代中更受关 \¥0 2019/144384 卩(:17€謂18/074353
8
注。
(4)输出
强学习机:
Figure imgf000010_0001
本发明以学习速度快、 泛化性能好的极限学习机 (ELM) 为弱学习机, 通 过设定合适的迭代次数以及阈值, 得到预测精度高的强学习机, 即航空发动机 启动过程排气温度预测模型。

Claims

\¥0 2019/144384 卩(:17€謂18/074353 9 权利要求书 1.一种航空发动机启动过程排气温度预测方法, 其特征在于, 对传感器采集到 的航空发动机启动过程地面试车数据, 采用基于密度的方法对数据进行异常点 的识别和处理, 采用二次指数平滑法对数据中的噪声或数据污染进行平滑或滤 波处理, 并对数据进行归一化处理使其转化为[0,1]范围内的数据; 基于信息融合的思想, 利用互信息法进行相关性分析, 计算各个参数与排 气温度的互信息函数值, 考虑到各个参数与排气温度相关性大小的差异及训练 预测模型的时间要求, 取互信息函数值最大的三个参数及排气温度作为预测模 型的输入参数; 设预处理后的数据为/¾¾, 其中 其中, /¾ 航空发动机性能参数数据, {¾} 相对应的时间序列, A ^为 参数个数, 《为样本个数; 对选择的参数进行相空间重构构建输入输出样本以充分显露时间序列数据 中隐含的信息; 设相关性分析后的数据为 /^ 1, 其中, \¥02019/144384 卩(:17(:\2018/074353 10 0^2 = [/¾<¾ ,尸〇<¾,尸<3<¾ , ParaA\ (7) 尸 ={xi}ni=vl = 1, 2,3,4 (8) 特别地, /^ 4=£(;7; 对于时间序列{¾};=1, 对其重构后的相空间为 其中, ¾ = [¾, ·¾/ ),…, ¾(/+(«-1>) ] , / = 1,2, , ; = « - ( - 1>· (10) 其中, 为嵌入维数, 是延迟时间, 分别用互信息法和〇1〇法进行求解, 根据相空间重构构造输入输出样本, 如表 1所示, 其中 为预测步长; 表 1基于相空间重构的输入输出数据 采用 AdaBoost.RT_ELM算法对航空发动机启动过程的排气温度进行预测,AdaBoost.RT_ELM具体算法如下:
(1)输入
Figure imgf000012_0002
\¥0 2019/144384 卩(:17€謂18/074353
11 样本和预测错误的样本
2)初始化
令初始迭代次数?=1
令第一次训练时, 训练样本权值分布为^(/)=1^/ =1,···^;
令初始误差率 =0
(3 )迭代过程
Figure imgf000013_0001
Step 3:设置 =<, a可以为 1,2或 3;
Step 4:更新样本权值 Z)f :
Figure imgf000013_0002
其中, ^是标准化因子; 通过调整各样本的权值, 即增加预测误差大的样 本的权值, 减少预测误差小的样本的权值, 使得误差大的样本在下一次迭代中 更受关注; 4)输出
强学习机: \¥0 2019/144384 卩(:17€謂18/074353
12
Figure imgf000014_0001
以学习速度快、 泛化性能好的极限学习机为弱学习机, 通过设定合适的迭 代次数以及阈值, 得到预测精度高的强学习机, 即航空发动机启动过程排气温 度预测模型。
PCT/CN2018/074353 2018-01-26 2018-01-26 一种航空发动机启动过程排气温度预测方法 WO2019144384A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
PCT/CN2018/074353 WO2019144384A1 (zh) 2018-01-26 2018-01-26 一种航空发动机启动过程排气温度预测方法
US16/629,893 US20200234165A1 (en) 2018-01-26 2018-01-26 Prediction method for aero-engine starting exhaust temperature

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/074353 WO2019144384A1 (zh) 2018-01-26 2018-01-26 一种航空发动机启动过程排气温度预测方法

Publications (1)

Publication Number Publication Date
WO2019144384A1 true WO2019144384A1 (zh) 2019-08-01

Family

ID=67395192

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/074353 WO2019144384A1 (zh) 2018-01-26 2018-01-26 一种航空发动机启动过程排气温度预测方法

Country Status (2)

Country Link
US (1) US20200234165A1 (zh)
WO (1) WO2019144384A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113158537A (zh) * 2021-01-18 2021-07-23 中国航发湖南动力机械研究所 基于lstm结合注意力机制的航空发动机气路故障诊断方法
CN118133690A (zh) * 2024-05-07 2024-06-04 深圳市安保医疗科技股份有限公司 样本构建方法、涡轮性能预测方法、装置、设备及介质

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111882130B (zh) * 2020-07-30 2022-01-11 浙江大学 一种基于生成路径聚类与Box-Cox变换的二噁英排放在线预测方法
CN112069699B (zh) * 2020-09-29 2022-08-23 上海工程技术大学 一种基于组合预测模型预测齿轮箱健康状态退化趋势的方法
CN113128612B (zh) * 2021-04-26 2022-11-29 国网河北省电力有限公司营销服务中心 电力数据中异常值的处理方法及终端设备
US20220343770A1 (en) * 2021-04-27 2022-10-27 Rockwell Collins, Inc. Machine-learned operating system and processor
CN113203471B (zh) * 2021-05-07 2022-08-02 国网山西省电力公司电力科学研究院 一种基于小波神经网络预测的变压器机械故障检测方法
CN113762471B (zh) * 2021-08-24 2023-07-28 浙江中辰城市应急服务管理有限公司 一种基于注意力机制与贝叶斯优化的相空间重构参数估计方法

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102562239A (zh) * 2011-11-13 2012-07-11 哈尔滨工业大学 一种航空发动机排气温度的监测方法
CN107131040A (zh) * 2016-02-26 2017-09-05 福特环球技术公司 用于确定排气温度的系统和方法

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102562239A (zh) * 2011-11-13 2012-07-11 哈尔滨工业大学 一种航空发动机排气温度的监测方法
CN107131040A (zh) * 2016-02-26 2017-09-05 福特环球技术公司 用于确定排气温度的系统和方法

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
FU, JINHUA ET AL.: "A Method of Modeling .. Exhaust Gas Temperature Baseline of Aero-engine", AERONAUTICAL COMPUTING TECHNIQUES, 30 November 2014 (2014-11-30), pages 49 - 54 *
ZHONG, SHISHENG ET AL.: "Convolution Sum Discrete Process Neural Network and Its Application in Aeroengine Exhausted Gas Temperature Prediction", ACTA AERONAUTICA ET ASTRONAUTICA SINICA, 25 March 2012 (2012-03-25), pages 438 - 444, XP055627683, ISSN: 1000-6893, Retrieved from the Internet <URL:https://www.semanticscholar.org/paper/Convolution-Sum-Discrete-Process-Neural-Network-and-Gang/9a2fe2e55f6f1f94bd37d43a1a0e090135f2c441#paper-header> *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113158537A (zh) * 2021-01-18 2021-07-23 中国航发湖南动力机械研究所 基于lstm结合注意力机制的航空发动机气路故障诊断方法
CN113158537B (zh) * 2021-01-18 2023-03-24 中国航发湖南动力机械研究所 基于lstm结合注意力机制的航空发动机气路故障诊断方法
CN118133690A (zh) * 2024-05-07 2024-06-04 深圳市安保医疗科技股份有限公司 样本构建方法、涡轮性能预测方法、装置、设备及介质

Also Published As

Publication number Publication date
US20200234165A1 (en) 2020-07-23

Similar Documents

Publication Publication Date Title
WO2019144384A1 (zh) 一种航空发动机启动过程排气温度预测方法
Liu et al. Prediction of remaining useful life of multi-stage aero-engine based on clustering and LSTM fusion
US10417353B2 (en) Method and system of dynamic model identification for monitoring and control of dynamic machines with variable structure or variable operation conditions
WO2021164250A1 (zh) 一种湍流场更新方法、装置及其相关设备
US20140278303A1 (en) Method and system of dynamic model identification for monitoring and control of dynamic machines with variable structure or variable operation conditions
CN108363844B (zh) 一种航空发动机启动过程排气温度预测方法
WO2023236393A1 (zh) 一种应用于燃气轮机组的故障预警方法、系统和装置
CN112231847A (zh) 转捩位置确定方法、装置、电子设备及存储介质
CN113361048B (zh) 燃气轮机叶片的可靠性预测方法、装置及电子设备
CN115017466B (zh) 碳载量确定方法、装置、电子设备及存储介质
Huang et al. Gas path deterioration observation based on stochastic dynamics for reliability assessment of aeroengines
CN115310359A (zh) 氮氧化物瞬态排放确定方法、装置、设备及介质
Shaowu et al. Prediction of exhaust gas temperature margin based on LSSVR
CN112182739A (zh) 一种飞行器结构非概率可信可靠性拓扑优化设计方法
CN114991923B (zh) 颗粒捕集器压差确定方法、装置、设备及介质
CN116595883B (zh) 数值反应堆实时在线系统状态修正方法
CN114996660B (zh) 碳载量预测方法、装置、电子设备及存储介质
Huang et al. Line sampling based fuzzy simulation coupled with adaptive Kriging for estimating failure possibility of simplified turbine disk
Zhou et al. An active learning variable-fidelity metamodeling approach for engineering design
CN116227367B (zh) 直接空冷系统的背压预测模型构建方法、预测方法及装置
CN114707098B (zh) 一种基于多源传感器状态距离的航空发动机性能退化状态评价方法
CN115222164B (zh) 一种基于经验耦合函数的水泵故障预测方法及系统
Beqiraj et al. Rotor37 Aerodynamic Optimization: A Machine Learning Approach
Huang Law of incipient separation for turbulent flows over airfoils as inferred by RANS
CN114859796A (zh) 一种车辆控制方法、电子设备及存储介质

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18902089

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18902089

Country of ref document: EP

Kind code of ref document: A1