| It is significantly economical valuable to the exploitation of oil fields,especially to the ones located on the sea and marginal oilfield that the methord of on-line measurement about the multiphase flow technology in oil wellhead.As it is very difficult to develop multiphase flowmeter due to the complexity of flow regimes and inconstancy of flow composition,so far there are few practicable multiphase flowmeters available.Focusing on the measurement of multiphase flow technology(the volume content of oil,gas and water),the paper provides a smart metering program based on artificial neural network(ANN).The program uses a multiphase flowmeter simulated oil well site as a experimental platform,applys the method that combined hardware and software to research of realizing device for three-phase measurement in oil wellhead based on ANN.Firstly,making use of the frequency,pressure and pressure difference factors which have significant impact on the three-phase measurement in oil wellhead to determine the neural network(NN) information expression.It presents and selects the historical datas of the factors as NN samples datas by multiphase flowmeter platform.Secondly,learning and training the selected datas by useing the NN toolbox of MATLAB in PC until it has to be the model,whose generalization ability and convergence rates are both better.Finally, useing the important parameters information(layers of NN,numbers of every layer,weights and threshold values of every neuron) of the received NN model to edit program,and then embedding it to micro controller unit hardware system.The values are displayed by the LED display circuit.A realizing device model of this design using AT89C55 MCU was developed.Result of the test shows that the errors are no more than 0.1 m~3/s,which are between the actual values of test samples and the measurement device output values.The experimental datas roughly demonstrate the program is effective and feasible.
|
PDF Full Text Request