A New Temperature Compensation method for flow Measuremen Employing FLNN

摘要

In a differential pressure flow measurement system,the outputs of the sensor are influenced by the environment temperature,which induces measurement errors.So temperature compensation for more accurate measurement is needed.The method of surface and curve fitting are usually applied in traditional digital compensation.In this paper digital compensation is also employed.But the mathematical model which applies Functional-link neural network(FLNN)is different from above digital compensation.The traditional digital compensation has fixed mathematical model and can’t change model according to different sensors compensation.FLNN has higher compensation accuracy and the model based on FLNN is flexible.FLNN is a single layer network without hidden-layer.FLNN transforms the low dimension to high dimension by input variable non-linearity function expanding as input of single feedback network.In this paper three common adopted function expanding forms are applied in simulation based on FLNN.The result of simulations show that employs Chebyshev polynomial in expanding function has higher compensation accuracy than other two function expanding forms.The max error of it is 0.75％ which can satisfy the demand of flow measurement.So Chebyshev polynomial function expanding form is applied in temperature compensation of differential pressure sensor.In fact,the flexibility of FLNN automatically compensates any variation of the sensor response occurring due to change in environmental conditions.It has a potential future in the field of measurement and control.