Skeletal model reduction based on local sensitivity analysis of time dependent systems is presented in which sensitivities are modeled by forced optimally time dependent (f-OTD) modes. The f-OTD factorizes the sensitivity coefficient matrix into a compressed format as the product of two skinny matrices, i.e. f-OTD modes and f-OTD coefficients. The modes create a low-dimensional, time dependent, orthonormal basis which capture the directions of the phase space associated with most dominant sensitivities. These directions highlight the instantaneous active species, and reaction paths. Evolution equations for the f-OTD modes and coefficients are derived, and the implementation of f-OTD for skeletal reduction is described. For demonstration, skeletal reduction is conducted of the constant pressure ethylene-air burning in a zero-dimensional reactor, and new reduced models are generated. The laminar flame speed, the ignition delay, and the extinction curve as predicted by the models are compared against some existing skeletal models in literature for the same detailed model. The results demonstrate the capability of f-OTD to eliminate unimportant reactions and species in a systematic, efficient and accurate manner.

介绍了基于时间相关系统局部灵敏度分析的骨架模型简化，其中灵敏度通过强制优化时间相关 (f-OTD) 模式建模。f-OTD 将灵敏度系数矩阵分解为两个瘦矩阵的乘积压缩格式，即f-OTD 模式和 f-OTD 系数。这些模式创建了一个低维的、时间相关的、正交的基础，它捕获了与大多数主要灵敏度相关联的相空间的方向。这些方向突出了瞬时活性物质和反应路径。推导出了 f-OTD 模式和系数的演化方程，并描述了 f-OTD 用于骨架减少的实现。为了演示，在零维反应器中对恒压乙烯-空气燃烧进行骨架还原，并生成新的还原模型。将模型预测的层流火焰速度、点火延迟和熄灭曲线与相同详细模型的文献中的一些现有骨架模型进行比较。