Intelligent Frequency Modulation In The Data Transmission Based On Neural Network

This paper discusses the measurement while drilling (MWD) system design in whole andintroduces the system in three parts. The first part is about the signal transmission indownhole. It mainly introduces the principle of the DSP as the core of the signal generatorsending signals. The downhole sensor will transmit some downhole logging parameters that ithas received to the DSP. Then the DSP will encode the signal in accordance with thepre-agreed communication protocol, and generates a continuous pressure wave throughcontrolling the brushless motor. The pressure wave will transmit to Inoue through mudsolution. The second part mainly introduces a variety of transmission media including thewired transmission mode and the wireless transmission mode. After comparing the featuresand performances of various transmission modes, the author selects the mud fluid as the finaltransmission medium. The mud fluid transmits the continuous wave generated by thedownhole signal transmitting portion to the Inoue. The Inoue sensor ultimately receives thesignal. The signal will be received by ground computer after being denoised. The downholelogging parameters can be obtained after the signal being decoded. And based on loggingparameters the computer can select the optimal control mode, and then send controlcommands to the downhole. A MWD system which can communicate with each other is thusconstituted.In the MWD system, noise is randomly generated and the noise frequency may be veryclose or even identical to the frequency of the signals used by the modulation signal.Therefore, with such a fixed signal frequency the final data reception error rate will be veryhigh or correct information cannot be received. In order to enhance the system’s ability toadapt to the environment, multi-frequency modulation can be used to complete thetransmission of the signal. The advantage of this approach is that it can adopt the mostfavorable data transmission mode intelligently and flexibly according to the changes of noiseand other environments.As the slurry continuous wave changes according to the measurement of drilling(MWD) date, the precision of error rate prediction is low and the process of transferring data will be affected by signals. Based on the BP neural network’s extensive mapping ability andchaos optimization algorithm’s global convergent ability, we structure a kind of improvedchaos optimization of BP neural network algorithm. This algorithm can avoid severalproblems, such as the convergent speed of BP neural network is slow and the BP neuralnetwork is easy to sink into local minimum. With the powerful ability of generalization andprediction, this kind of algorithm can also be used to predict the data transmission error rate inslurry continuous wave. Under the condition of small samples, The author create a model ofdata transmission in slurry continuous wave, which is based on improved chaos optimizationof BP neural network. Simulate experiment has tested this algorithm’s feasibility andeffectiveness.