| The measurement is the important mesns of mankind to know the physical world then reform it.Transducer is the instrucment that feels the examine signal and it is hoped that it can output the useful signal according to the certain regulation. As itself factor of transducer ,the output of transducer present not line. And under most condition ,because of affect by the electromagnetic wave,noise and the external world environment ,not line of transducer becomes more aomplicated.In modern measurement system ,the relation of input-output widespread not line,and beause the measurement is effected by temperature and noise,not line becomes more complicated.Using the traditional method to estabish synthesize correct mathematics model is more difficult. But the network have the ability to pemit wrong and of associationn of thoughts to remember,and have strong ability of adaptation ang not line reflect .It is easy to resolve the problem of not line approaches.Therefore using the network to establish characteristic model and error revied model of measurement system is a kind of ideal choice.In this paper ,the Neural Network is leaded into measurement system ,and using the Neural Network to establish synthetic correction model of the measurenment system to improve the accuracy of the system. In the paper, we use the Backpropagation Neural Network that has very strong ability of reflection ang shoot and adopts backpropagation arithmetic to learn to to establish synthetic correction model of the measurenment system and use quick backpropagation arithmetic to train the network.By simulation, get valid network model of the measure system.The main works are as follows.1. Modelling of the Pressure effect on Debsity in 2-Propanone under a Broad Range of Temperature2. Modelling of the Pressure effect on Acoustic Speed in 2-Propanone under a Broad Range of Temperatur3. Modelling of Ultrasonic Not-inserting sensing system of the Hydrochloric Acid...
|
PDF Full Text Request