State estimation is DTS's foundation software, bad data processing is stateestimation's key component. We disposal DTS's original data, through stateestimation theory to get higher precise data. According to Huber estimation,deduce the concept of weight factor, applicate weight factor into bad dataidentification with directly use non quadratic identification criterioniteration formula. Decrease iteration times, acceleration speed withincorrection weight factor by residual variation. The test simulation in HeNanDTS system shows that this method is feasible.
|