Modeling Packages And General Dynamic Compensation Filter Developed

Sensors are core parts of instruments and measurement systems and the first sectors of process control system. In test and measurement system, if the dynamic characteristic of sensors can not meet the requirement, sensors will cause dynamic errors. So it is important for measurement system to improve the dynamic characteristic of sensors. The study on dynamic modeling and compensation of sensors is given.Three methods of sensor modeling, general least square (GLS), Walsh function and neural network, are proposed in this thesis. And accurate models are obtained. According to the idea of verse model, sensor's dynamic compensation filter is designed. At the same time, certain sensor is modeled and compensated according to its calibrating data. Good results are obtained. Soft box of sensor modeling, which is programmed in matlab, is given to users. In addition, the application of BP neural network on sensor modeling is studied to some extent. Some improved algorithms are proposed to improve the applicability of BP neural network on sensor modeling.When it comes to dynamic compensation filter, the design method of general dynamic compensation filter is presented from digital and analog aspects. The theoretical feasibility is verified. Finally, hardware circuit of dynamic compensation filter is designed.