Spatial gradient measurement through length scale estimation for the tracking of a gaseous source

Material release, deliberate or accidental, into the atmosphere can pose a significant hazard. Knowledge of the location of the source can be used to contain the damage by stopping the threat from additional material release. However, in order to minimize the effects of the material already released, an accurate estimate of the released plume must be available. This work uses such a model based approach to predict the state of the contaminant within a spatial domain. A mobile agent provides sensor measurements in order to efficiently reconstruct the process state. Sensor guidance is based on Lyapunov methods which drives the sensor to areas of higher state estimation error. As part of this guidance, the spatial gradient of the estimation error is used, which must be computed from sensor measurements. In flows with turbulence and eddies, care must be taken when determining the local gradient though the appropriate length scale. This will minimize the effects of the stochastic components of the flow on the measurements. With a model based estimate of the plume, the appropriate length scale can be predicted from sensor measurements. This work investigates the use of a measured spatial gradient based on an estimated length scale at the sensor location and compares the results with ideal spatial gradient data.