The Algorithm Study of Sensor Compensation in MWD Instrument Based on Genetic Elman Neural Network
Li-li Ju, Xiufang Wang, Sai Ma, Wei Chun-ming · 2010
In order to improve the measurement precision and stability of MWD Instrument, we create Elman neural network model and utilize self-adaptive genetic algorithm to optimize weights threshold value of the right of Elman network, which overcomes the disadvantages of traditional method, such as training for a long time, easy to fall into local optimal solution. Simulation results show that the error accuracy increases 3 orders of magnitude, compared with Elman network, the compensation effect is very stable.