針對四輪車輪馬達直接驅動之電動車,本文利用以輪胎旋轉動態為基礎的模型,來取代成本較高的硬體冗贅配置的偵測機制。配合遞迴最小平方法可識別等效的縱向輪胎剛性與軸承阻尼係數,進行不同等級的車輪故障偵測與隔離。如果偵測出某一車輪馬達是屬於部分故障的話,只需針對控制分配後的車輪馬達力矩,進行軸承阻尼力矩的前饋補償;如果偵測出某一車輪馬達是屬於完全故障的話,則須進行控制分配重組態,利用其他三個正常的車輪馬達來重新進行力矩分配。本技術所提出的方法,可在某一車輪馬達出現部分或全部故障的情形下,達到主動式容錯控制的功能,以保持系統的穩定性與維持一定的操控性能。
For the electric vehicle with four direct-driven wheel motors, this technology utilizes the tire rotational dynamics as the model to replace the detection mechanism using redundant hardware with higher cost. Recursive least square method is used to identify the equivalent longitudinal tire stiffness and bearing damping coefficient for different levels of wheel fault detection and isolation. If one wheel with partial fault is detected, feedforward compensation of the bearing damping torque is added to the wheel motor torque distributed by the control allocation. If the one wheel with complete fault is detected, control allocation is necessary to be reconfigured. Other remaining normal wheels will be used for torque distribution. This technology can achieve active fault tolerant control to maintain the system stability and a certain maneuverability with the presentence of partial or complete fault of the wheel motor.