Insect wings act not simply as actuators, generating the forces necessary for flight, but also act as sensory structures that provide rapid feedback for stable flight control. Wing structural mechanics and motions determine wing strains and therefore neural signals conveyed by sensory structures. Prior studies of sensing in flapping wings relied on analytical models based on Euler-Lagrange equations. Although these approaches are computationally tractable, they preclude more complex structures and kinematics associated with real wing motions. Here, we developed a finite element model that allows us to investigate sensing in the context of more realistic wing structural mechanics and movements. We developed a simplified model that shares several features with wings of the hawkmoth Manduca sexta and our prior analytic models of a flapping plate subject to rotation orthogonal to the flapping motion. The model simulates the spatial and temporal pattern of strain over the wing. That strain is then encoded by a neural-inspired transform which includes a linear filter and non-linear threshold to predict spatial and temporal patterns of neural spiking. We then used that spiking pattern, along with a compressive sensing algorithm to show that we can detect body rotations of comparable magnitude to those experienced by Manduca during flight with greater than 95% accuracy with only five sensors. This framework allows us to incorporate a variety of wing structural properties and kinematics, and to investigate sensing in the context of biologically realistic wings.