齿面微点蚀演变对直齿轮啮合功耗的影响规律研究

This research, targeting at spur gear, is aimed at revealing the influence law of micropitting evolution on gear-mesh power loss theoretically and experimentally. A calculation method of gear-mesh power loss with micropitting is proposed and on-line monitoring gear micropitting and gear-mesh power loss is achieved. Firstly, a 3D numerical modeling method integrating partial film EHL theory, multi-axial fatigue theory and extended finite element method (XFEM) is used to analyze the evolution of micropitting over time. Then, a coupling analysis method of EHL and dynamic load is adopted to calculate gear-mesh power loss under the micropitting condition, thus analyzing change law of the mesh power loss with the micropitting evolution. Moreover, bubbles in OLVF ferrograms is recognized based on superpixel and deep converlutional networks, and wear debris are segmented based on frame difference and Watershed, so that wear debris size distribution of micropitting is extracted. Accordingly, on-line monitoring micropitting and gear-mesh power loss is achieved, and then the evolution curves of micropitting and wear debris size distribution over time is obtained to verify the theoretical results. The results become a new technology support for proactively controlling micropitting to promote the tribology design of gears, and provide a new technique for ensuring low carbon, safety in services of high-power gear transmission systems.

本项目以直齿轮为对象，提出微点蚀条件下齿轮啮合功耗的计算方法，在线监测齿面微点蚀和齿轮啮合功耗，揭示微点蚀演变对齿轮啮合功耗的影响规律。首先，结合部分弹流润滑理论、多轴疲劳理论和扩展有限元法，提出三维微点蚀齿面建模方法，分析齿面微点蚀随时间的演变规律。其次，提出齿面微点蚀条件下的齿轮弹流润滑和啮合动载荷耦合分析方法，实现微点蚀条件下的齿轮啮合功耗计算，分析齿轮啮合功耗随齿面微点蚀演变的变化规律。然后，提出基于超像素和深度卷积神经网络的气泡识别算法，以及基于帧间差分和分水岭的磨粒分割算法，实现气泡干扰下的微点蚀磨粒粒度特征提取。在此基础上实现齿面微点蚀和齿轮啮合功耗的在线监测，获取微点蚀磨粒粒度分布和齿轮啮合功耗的演变曲线，验证理论分析结果。相关成果为微点蚀的主动控制提供新的技术支撑，推进齿轮摩擦学设计技术的发展，另一方面，为齿轮健康监测提供新的技术路径，促进大功率齿轮系统的安全、低碳运行。

本项目以直齿轮为对象，开展齿面微点蚀演变的数值模拟分析和微点蚀磨粒在线监测技术研究，揭示微点蚀演变对齿轮啮合功耗的影响规律。考虑了高斯表面粗糙度峰值数的影响，运用流变学方法计算最小粗糙度峰值，模拟微点蚀裂纹萌生过程，预测齿面微点蚀区域；考虑齿轮的接触特性和弹流润滑的影响，提出了一种微点蚀裂纹扩展模拟分析方法，分析了齿轮运行条件对生成磨粒尺寸的影响；针对透射光谱片气泡干扰问题，提出了基于UDCT的气泡抑制算法，实现了谱片气泡的抑制；针对反射光谱片气泡干扰问题，提出了基于深度学习的磨粒分割算法，实现了气泡干扰条件下的反射光磨粒分割，满足齿轮微点蚀在线监测的需求；以在线图像可视铁谱2D高梯度磁场为基础，提出了基于高梯度静磁场感应式磨粒传感方法，实现了25 μm铁磁性磨粒监测；为进一步提升磨粒监测效能满足微点蚀磨粒监测的要求，提出了基于旋转对称高梯度磁场的磨粒传感方法，实现13 μm铁磁性磨粒的在线监测，满足了微点蚀的在线监测需求；然后研究磨粒感应信号的混叠问题，分析了磨粒混叠感应信号的波形特征，为磨粒量化分析奠定了基础；最后针对大流润滑油磨粒的监测需求，提出了多激励高梯度磁场感应式磨粒传感方法，为风电齿轮箱等装备的微点蚀在线监测奠定基础。