动态参数与服役行为映射关系分析算法

本数据集主要面向物理知识与运行数据驱动的重大装备异常检测与故障诊断研究，针对直升机三器两轴串联式无冗余结构，开展直升机动态响应参数与服役行为关联关系分析。通过深度编码器和深度卷积神经网络组成的联合模型，厘清运行状态识别与工况变化的关联关系，为阐明工况变化对状态监测数据及信息的影响机制提供前提；通过压缩感知与元强化学习，实现工况变化免疫特征参数的提取，为进一步开展智能诊断与预测提供研究基础；通过改进相空间曲变方法实现变工况下健康指数生成与优选，为进一步开展高准确度的智能诊断与预测提供样本工具。

This dataset is primarily intended for research on anomaly detection and fault diagnosis of large-scale equipment driven by physical knowledge and operational data. Focusing on the series-connected non-redundant three-device two-shaft structure of helicopters, it conducts analysis on the correlation between helicopter dynamic response parameters and service behavior. A joint model consisting of a deep encoder and a deep convolutional neural network is employed to clarify the association between operating state recognition and working condition variations, providing a prerequisite for elucidating the influence mechanism of working condition changes on condition monitoring data and information. Through compressed sensing and meta-reinforcement learning, the extraction of feature parameters immune to working condition variations is achieved, laying a research foundation for further intelligent diagnosis and prediction. An improved phase space warping method is utilized to generate and optimize health indicators under variable working conditions, providing sample tools for subsequent high-accuracy intelligent diagnosis and prediction.