基于改进无模型自适应算法的涡扇发动机限制保护控制方法

针对发动机限制保护控制问题,提出了一种基于数据驱动的改进无模型自适应限制保护控制方法。以主回路闭环系统为设计对象,通过在常规无模型自适应控制律的准则函数中增加误差变化率,使超限保护系统具有更快的响应速度,给出了改进后系统的稳定性条件,并采用临界比例度法进行了控制器参数整定。仿真结果表明:在不同工作点,所设计的改进无模型自适应限制保护控制器相比常规无模型自适应限制保护控制器,最大超限平均减小2.97%,响应时间平均缩短2.30s,实现了发动机超限后的迅速回调,保证了发动机系统的安全运行。     

Framework and development of data-driven physics based model with application in dimensional accuracy prediction in pocket milling

In the manufacturing of thin wall components for aerospace industry, apart from the side wall contour error, the Remaining Bottom Thickness Error (RBTE) for the thin-wall pocket component (e.g. rocket shell) is of the same importance but overlooked in current research. If the RBTE reduces by 30%, the weight reduction of the entire component will reach up to tens of kilograms while improving the dynamic balance performance of the large component. Current RBTE control requires the off-process measurement of limited discrete points on the component bottom to provide the reference value for compensation. This leads to incompleteness in the remaining bottom thickness control and redundant measurement in manufacturing. In this paper, the framework of data-driven physics based model is proposed and developed for the real-time prediction of critical quality for large components, which enables accurate prediction and compensation of RBTE value for the thin wall components. The physics based model considers the primary root cause, in terms of tool deflection and clamping stiffness induced Axial Material Removal Thickness (AMRT) variation, for the RBTE formation. And to incorporate the dynamic and inherent coupling of the complicated manufacturing system, the multi-feature fusion and machine learning algorithm, i.e. kernel Principal Component Analysis (kPCA) and kernel Support Vector Regression (kSVR), are incorporated with the physics based model. Therefore, the proposed data-driven physics based model combines both process mechanism and the system disturbance to achieve better prediction accuracy. The final verification experiment is implemented to validate the effectiveness of the proposed method for dimensional accuracy prediction in pocket milling, and the prediction accuracy of AMRT achieves 0.014 mm and 0.019 mm for straight and corner milling, respectively.     

火箭发动机涡轮泵故障检测及预测算法研究

涡轮泵作为液体火箭发动机的核心部件,恶劣的工作环境和极高的转速使其易发生组件断裂、烧蚀等问题。为了对液体火箭发动机的涡轮泵进行健康管理,提出针对某型液体火箭发动机涡轮泵的数据驱动故障检测、故障预测及健康状态评估方法。在某型液体火箭发动机试车数据集上,通过对涡轮泵轴、径、切向振动数据进行对应的时域、频域特征处理后,送入训练好的ResNet网络、自主设计的图像特征识别算法以及退化模式线性回归模型,分别实现了对该型液体火箭发动机涡轮泵的故障检测、预测及健康状态评估,具有较高的准确性。     

基于粒子滤波的结构疲劳裂纹扩展动态贝叶斯推断方法

准确预测结构的疲劳裂纹扩展过程是开展飞机单机寿命监控与剩余寿命估算的基础。提出一种基于动态贝叶斯网络的结构疲劳裂纹扩展预测方法,结合疲劳裂纹扩展的先验知识与后验知识来准确地推断裂纹长度;研究粒子滤波算法中不同粒子数对动态贝叶斯网络推断精度的影响规律;通过对单孔板结构与耳片连接结构件在随机载荷谱下进行裂纹扩展研究。结果表明：动态贝叶斯网络方法可以对复杂结构的疲劳裂纹扩展进行准确预测,预测精度相对于传统方法提高50% 以上。     

In-process adaptive milling for large-scale assembly interfaces of a vertical tail driven by real-time vibration data

Assembly interfaces, the joint surfaces between the vertical tail and rear fuselage of a large aircraft, are thin-wall components. Their machining quality are seriously restricted by the machining vibration. To address this problem, an in-process adaptive milling method is proposed for the large-scale assembly interface driven by real-time machining vibration data. Within this context, the milling operation is first divided into several process steps, and the machining vibration data in each pro...     

Improving EGT sensing data anomaly detection of aircraft auxiliary power unit

The reliability of the on-wing aircraft Auxiliary Power Unit (APU) decides the cost and the comfort of flight to a large degree. The most important function of APU is to help start main engines by providing compressed air. Especially on the condition of sudden shutdown in the air, APU can offer additional thrust for landing. Therefore, its condition monitoring has drawn much attention from the academic and industrial field. Among the on-wing sensing data which can reflect its condition, Exhaust Gas Temperature (EGT) is one of the most important parameters. To ensure the reliability of EGT, one kind of data-driven anomaly detection framework for EGT sensing data is proposed based on the Gaussian Process Regression and Kernel Principal Component Analysis. The situations of one-dimensional and two-dimensional input data for EGT anomaly detection are considered, respectively. The cross-validation experiments are carried out by utilizing the real condition data of APU, which are provided by China Southern Airlines Company Limited Shenyang Maintenance Base. The anomalous stuck condition of EGT sensing data is also detected. Experimental results show that the proposed EGT sensing data anomaly detection method can achieve better performance of false positive ratio, false negative ratio and accuracy.     

Identifying and managing risks of AI-driven operations: A case study of automatic speech recognition for improving air traffic safety

In this work, the primary focus is to identify potential technical risks of Artificial Intelligence (AI)-driven operations for the safety monitoring of the air traffic from the perspective of speech communication by studying the representative case and evaluating user experience. The case study is performed to evaluate the AI-driven techniques and applications using objective metrics, in which several risks and technical facts are obtained to direct future research. Considering the safety–critical specificities of the air traffic control system, a comprehensive subjective evaluation is conducted to collect user experience by a well-designed anonymous questionnaire and a face-to-face interview. In this procedure, the potential risks obtained from the case study are confirmed, and the impacts on human working are considered. Both the case study and the evaluation of user experience provide compatible results and conclusions: (A) the proposed solution is promising to improve the traffic safety and reduce the workload by detecting potential risks in advance; (B) the AI-driven techniques and whole diagram are suggested to be enhanced to eliminate the possible distraction to the attention of air traffic controllers. Finally, a variety of strategies and approaches are discussed to explore their capability to advance the proposed solution to industrial practices.     

Data-driven model-free adaptive attitude control of partially constrained combined spacecraft with external disturbances and input saturation

This study presents an improved data-driven Model-Free Adaptive Control(MFAC)strategy for attitude stabilization of a partially constrained combined spacecraft with external disturbances and input saturation. First, a novel dynamic linearization data model for the partially constrained combined spacecraft with external disturbances is established. The generalized disturbances composed of external disturbances and dynamic linearization errors are then reconstructed by a Discrete Extended State Observer(DESO). With the dynamic linearization data model and reconstructed information, a DESO-MFAC strategy for the combined spacecraft is proposed based only on input and output data. Next, the input saturation is overcome by introducing an antiwindup compensator. Finally, numerical simulations are carried out to demonstrate the effectiveness and feasibility of the proposed controller when the dynamic properties of the partially constrained combined spacecraft are completely unknown.     

Modeling satellite battery aging for an operational satellite simulator

During the satellite’s operations, simulation tools perform an important role in ensuring the space mission success. In this sense, the models implemented in the context of an operational satellite simulator that enables analysis of health status and maintenance during operations shall reflect the current satellite behavior with high fidelity. Moreover, it is complicated to obtain all analytical models of a satellite’s disciplines, considering its aging. This paper proposes an Artificial Neural Network (ANN) to reproduce the battery voltage behavior of a large sun-synchronous remote sensing satellite, the CBERS-4, currently in operation. Using the genetic algorithm to find the best network architecture of ANN, the neural model for this application presented an error of less than 1%, demonstrating its feasibility to obtain a high fidelity model for an operational simulator enabling extend analyses. The paper addresses advanced techniques aligned with the space industry’s future, increasing the ability to analyze a large amount of data and improve the space system’s operation.     

数字化车间虚拟监控系统研究

传统的虚拟监控系统管理层与执行层信息交互能力不足,不能满足当前车间监控需求,基于虚拟现实技术与信息集成技术,建立车间状态监控系统,实现整个车间内人员、物料、设备等位置及状况的实时监控.系统以三维制造资源模型为基础构建数据逻辑模型;采用XML规范建立系统模块间的数据接口;利用实时状态数据驱动三维虚拟模型;利用Unity3D技术实现虚拟场景的渲染与实时人机交互,最后通过实例验证了系统的有效性.